Overview

The Web API is a component which makes it possible for external systems to access and manipulate data stored in an instance of DHIS2. More precisely, it provides a programmatic interface to a wide range of exposed data and service methods for applications such as third-party software clients, web portals and internal DHIS2 modules.

Introduction

The Web API adheres to many of the principles behind the REST architectural style. To mention some important ones:

1. The fundamental building blocks are referred to as resources. A resource can be anything exposed to the Web, from a document to a business process - anything a client might want to interact with. The information aspects of a resource can be retrieved or exchanged through resource representations. A representation is a view of a resource's state at any given time. For instance, the visualizations resource in DHIS2 represents visualizations of aggregated data for a certain set of parameters. This resource can be retrieved in a variety of representation formats including JSON and CSV.
2. All resources can be uniquely identified by a URI (also referred to as URL). All resources have a default representation. You can indicate that you are interested in a specific representation by supplying an Accept HTTP header, a file extension or a format query parameter. So in order to retrieve a CSV representation of an analytics data response you can supply an Accept: application/csv header or append .csv or ?format=csv to your request URL.
3. Interactions with the API requires the correct use of HTTP methods or verbs. This implies that for a resource you must issue a GET request when you want to retrieve it, POST request when you want to create one, PUT when you want to update it and DELETE when you want to remove it.

Authentication

The DHIS2 Web API supports three protocols for authentication:

• Basic Authentication
• Personal Access Tokens (PAT)
• OAuth 2

You can verify and get information about the currently authenticated user by making a GET request to the following URL:

/api/33/me

And more information about authorities (and if a user has a certain authority) by using the endpoints:

/api/33/me/authorities
/api/33/me/authorities/ALL

Basic Authentication

The DHIS2 Web API supports Basic authentication. Basic authentication is a technique for clients to send login credentials over HTTP to a web server. Technically speaking, the username is appended with a colon and the password, Base64-encoded, prefixed Basic and supplied as the value of the Authorization HTTP header. More formally that is:

Authorization: Basic base64encode(username:password)

Most network-aware development environments provide support for Basic authentication, such as Apache HttpClient and Spring RestTemplate. An important note is that this authentication scheme provides no security since the username and password are sent in plain text and can be easily observed by an attacker. Using Basic is recommended only if the server is using SSL/TLS (HTTPS) to encrypt communication with clients. Consider this a hard requirement in order to provide secure interactions with the Web API.

Two-factor authentication

DHIS2 supports two-factor authentication. This can be enabled per user. When enabled, users will be asked to enter a 2FA code when logging in. You can read more about 2FA here.

Personal Access Token

Personal access tokens (PATs) are an alternative to using passwords for authentication to DHIS2 when using the API.

PATs can be a more secure alternative to HTTP Basic Authentication, and should be your preferred choice when creating a new app/script etc.

HTTP Basic Authentication is considered insecure because, among other things, it sends your username and password in clear text. It may be deprecated in future DHIS2 versions or made opt-in, meaning that basic authentication would need to be explicitly enabled in the configuration.

Important security concerns!

Your PATs will automatically inherit all the permissions and authorizations your user has. It is therefore extremely important that you limit the access granted to your token depending on how you intend to use it, see Configuring your token.

If you only want the token to have access to a narrow and specific part of the server, it is advised to rather create a new special user that you assign only the roles/authorities you want it to have access to.

Creating a token

To create a new PAT, you have two choices: * A. Create a token in the UI on your account's profile page. * B. Create a token via the API.

A. Creating a token on the account's page

Log in with your username and password, go to your profile page (Click top right corner, and chose "Edit profile" from the dropdown). On your user profile page, choose "Personal access tokens" from the left side menu. You should now be on the "Manage personal access tokens" page and see the text: "You don't have any active personal access tokens". Click "Generate new token" to make a new token. A "Generate new token" popup will be shown and present you with two choices:

1. Server/script context:

"This type is used for integrations and scripts that won't be accessed by a browser".

If you plan to use the token in an application, a script or similar, this type should be your choice.

2. Browser context:

"This type us used for applications, like public portals, that will be accessed with a web browser".

If you need to link to DHIS2 on a webpage, or e.g. embed in an iframe, this is probably the type of token you want.

Configuring your token

After choosing what token type you want, you can configure different access constraints on your token. By constraint, we mean how to limit and narrow down how your token can be used. This can be of crucial importance if you plan on using the token in a public environment, e.g. on a public dashboard on another site, embedded in an iframe. Since tokens always have the same access/authorities that your user currently has, taking special care is needed if you intend to use it in any environment you don't have 100% control over.

NB: If anyone else gets their hands on your token, they can do anything your user can do. It is not possible to distinguish between actions performed using the token and other actions performed by your user.

Important: It is strongly advised that you create a separate unique user with only the roles/authorities you want the token to have if you plan on using PAT tokens in a non-secure and/or public environment, e.g. on a PC or server, you don't have 100% control over, or "embedded" in a webpage on another server.

The different constraint types are as follows:

• Expiry time
• Allowed UP addresses
• Allowed HTTP methods
• Allowed HTTP referrers

Expiry time

Expiry time simply sets for how long you want your token to be usable, the default is 30 days. After the expiry time, the token will simply return a 401 (Unauthorized) message. You can set any expiry time you want, but it is strongly advised that you set an expiry time that is reasonable for your use case.

Allowed IP addresses

This is a comma-separated list of IP addresses you want to limit where the token requests can come from.

Important: IP address validation relies on the X-Forwarded-For header, which can be spoofed. For security, make sure a load balancer or reverse proxy overwrites this header.

Allowed HTTP methods

A comma-separated list of HTTP methods you want your token to be able to use. If you only need your token to view data, not modify or delete, selecting only the GET HTTP method makes sense.

Allowed HTTP referrers

HTTP referer is a header added to the request, when you click on a link, this says which site/page you were on when you clicked the link. Read more about the HTTP referer header here: https://en.wikipedia.org/wiki/HTTP_referer

This can be used to limit the use of a "public" token embedded on another page on another site. Making sure that the referer header match the site hostname in should come from, can help avoid abuse of the token, e.g. if someone posts it on a public forum.

Important: this is not a security feature. The referer header can easily be spoofed. This setting is intended to discourage unauthorized third-party developers from connecting to public access instances.

Saving your token:

When you are done configuring your token, you can save it by clicking the "Generate new token" button, on the bottom right of the pop-up. When doing so the token will be saved and a secret token key will be generated on the server. The new secret token key will be shown on the bottom of the PAT token list with a green background, and the text "Newly created token". The secret token key will look similar to this:

d2pat_5xVA12xyUbWNedQxy4ohH77WlxRGVvZZ1151814092

B. Creating a token via the API

Example of how to create a new Personal Access Token with the API:

POST https://play.dhis2.org/dev/api/apiToken
Content-Type: application/json
Authorization: Basic admin district

{}

This will return a response containing a token similar to this:

{
  "httpStatus": "Created",
  "httpStatusCode": 201,
  "status": "OK",
  "response": {
     "responseType": "ApiTokenCreationResponse",
     "key": "d2pat_5xVA12xyUbWNedQxy4ohH77WlxRGVvZZ1151814092",
     "uid": "jJYrtIVP7qU",
     "klass": "org.hisp.dhis.security.apikey.ApiToken",
     "errorReports": []
  }
}

Important: The token key will only be shown once here in this response. You need to copy and save this is in a secure place for use later!

The token itself consists of three parts: 1. Prefix: (d2pat_) indicates what type of token this is. 2. Random bytes Base64 encoded: (5xVA12xyUbWNedQxy4ohH77WlxRGVvZZ) 3. CRC32 checksum: (1151814092) the checksum part is padded with 0 so that it always stays ten characters long.

Configure your token via the API:

To change any of the constraints on your token, you can issue the following HTTP API request.

NB: Only the constraints are possible to modify after the token is created!

PUT https://play.dhis2.org/dev/api/apiToken/jJYrtIVP7qU
Content-Type: application/json
Authorization: Basic admin district

{
  "version": 1,
  "type": "PERSONAL_ACCESS_TOKEN",
  "expire": 163465349603200,
  "attributes": [
      {
        "type": "IpAllowedList",
        "allowedIps": ["192.168.0.1"]
      },
      {
        "type": "MethodAllowedList",
        "allowedMethods": ["GET"]
      }
  ]
}

Using your Personal Access Token

To issue a request with your newly created token, use the Authorization header accordingly. The Authorization header format is:

Authorization: ApiToken [YOUR_SECRET_API_TOKEN_KEY]

GET https://play.dhis2.org/dev/api/apiToken/jJYrtIVP7qU
Content-Type: application/json
Authorization: ApiToken d2pat_5xVA12xyUbWNedQxy4ohH77WlxRGVvZZ1151814092

Deleting your Personal Access Token

You can delete your PATs either in the UI on your profile page where you created it, or via the API like this:

DELETE https://play.dhis2.org/dev/api/apiToken/jJYrtIVP7qU
Content-Type: application/json
Authorization: ApiToken d2pat_5xVA12xyUbWNedQxy4ohH77WlxRGVvZZ1151814092

OAuth2

DHIS2 supports the OAuth2 authentication protocol. OAuth2 is an open standard for authorization which allows third-party clients to connect on behalf of a DHIS2 user and get a reusable bearer token for subsequent requests to the Web API. DHIS2 does not support fine-grained OAuth2 roles but rather provides applications access based on user roles of the DHIS2 user.

Each client for which you want to allow OAuth 2 authentication must be registered in DHIS2. To add a new OAuth2 client go to Apps > Settings > OAuth2 Clients in the user interface, click Add new and enter the desired client name and the grant types.

Adding a client using the Web API

An OAuth2 client can be added through the Web API. As an example, we can send a payload like this:

{
  "name": "OAuth2 Demo Client",
  "cid": "demo",
  "secret": "1e6db50c-0fee-11e5-98d0-3c15c2c6caf6",
  "grantTypes": [
    "password",
    "refresh_token",
    "authorization_code"
  ],
  "redirectUris": [
    "http://www.example.org"
  ]
}

The payload can be sent with the following command:

SERVER="https://play.dhis2.org/dev"
curl -X POST -H "Content-Type: application/json" -d @client.json
  -u admin:district "$SERVER/api/oAuth2Clients"

We will use this client as the basis for our next grant type examples.

Grant type password

The simplest of all grant types is the password grant type. This grant type is similar to basic authentication in the sense that it requires the client to collect the user's username and password. As an example we can use our demo server:

SERVER="https://play.dhis2.org/dev"
SECRET="1e6db50c-0fee-11e5-98d0-3c15c2c6caf6"

curl -X POST -H "Accept: application/json" -u demo:$SECRET "$SERVER/uaa/oauth/token"
  -d grant_type=password -d username=admin -d password=district

This will give you a response similar to this:

{
  "expires_in": 43175,
  "scope": "ALL",
  "access_token": "07fc551c-806c-41a4-9a8c-10658bd15435",
  "refresh_token": "a4e4de45-4743-481d-9345-2cfe34732fcc",
  "token_type": "bearer"
}

For now, we will concentrate on the access_token, which is what we will use as our authentication (bearer) token. As an example, we will get all data elements using our token:

SERVER="https://play.dhis2.org/dev"
curl -H "Authorization: Bearer 07fc551c-806c-41a4-9a8c-10658bd15435" "$SERVER/api/33/dataElements.json"

Grant type refresh_token

In general the access tokens have limited validity. You can have a look at the expires_in property of the response in the previous example to understand when a token expires. To get a fresh access_token you can make another round trip to the server and use refresh_token which allows you to get an updated token without needing to ask for the user credentials one more time.

SERVER="https://play.dhis2.org/dev"
SECRET="1e6db50c-0fee-11e5-98d0-3c15c2c6caf6"
REFRESH_TOKEN="a4e4de45-4743-481d-9345-2cfe34732fcc"

curl -X POST -H "Accept: application/json" -u demo:$SECRET "$SERVER/uaa/oauth/token"
  -d "grant_type=refresh_token" -d "refresh_token=$REFRESH_TOKEN"

The response will be exactly the same as when you get a token to start with.

Grant type authorization_code

Authorized code grant type is the recommended approach if you don't want to store the user credentials externally. It allows DHIS2 to collect the username/password directly from the user instead of the client collecting them and then authenticating on behalf of the user. Please be aware that this approach uses the redirectUris part of the client payload.

Step 1: Visit the following URL using a web browser. If you have more than one redirect URIs, you might want to add &redirect_uri=http://www.example.org to the URL:

SERVER="https://play.dhis2.org/dev"
$SERVER/uaa/oauth/authorize?client_id=demo&response_type=code

Step 2: After the user has successfully logged in and accepted your client access, it will redirect back to your redirect uri like this:

http://www.example.org/?code=XYZ

Step 3: This step is similar to what we did in the password grant type, using the given code, we will now ask for an access token:

SERVER="https://play.dhis2.org/dev"
SECRET="1e6db50c-0fee-11e5-98d0-3c15c2c6caf6"

curl -X POST -u demo:$SECRET -H "Accept: application/json" $SERVER/uaa/oauth/token
-d "grant_type=authorization_code" -d "code=XYZ"

Error and info messages

The Web API uses a consistent format for all error/warning and informational messages:

{
  "httpStatus": "Forbidden",
  "message": "You don't have the proper permissions to read objects of this type.",
  "httpStatusCode": 403,
  "status": "ERROR"
}

Here we can see from the message that the user tried to access a resource I did not have access to. It uses the http status code 403, the HTTP status message forbidden and a descriptive message.

Date and period format

Throughout the Web API, we refer to dates and periods. The date format is:

yyyy-MM-dd

For instance, if you want to express March 20, 2014, you must use 2014-03-20.

The period format is described in the following table (also available on the API endpoint /api/periodTypes)

Relative Periods

In some parts of the API, like for the analytics resource, you can utilize relative periods in addition to fixed periods (defined above). The relative periods are relative to the current date and allow e.g. for creating dynamic reports. The available relative period values are:

THIS_WEEK, LAST_WEEK, LAST_4_WEEKS, LAST_12_WEEKS, LAST_52_WEEKS,
THIS_BIWEEK, LAST_BIWEEK, LAST_4_BIWEEKS,
THIS_MONTH, LAST_MONTH, THIS_BIMONTH, LAST_BIMONTH, THIS_QUARTER, LAST_QUARTER,
THIS_SIX_MONTH, LAST_SIX_MONTH, MONTHS_THIS_YEAR, QUARTERS_THIS_YEAR,
THIS_YEAR, MONTHS_LAST_YEAR, QUARTERS_LAST_YEAR, LAST_YEAR, LAST_5_YEARS, LAST_10_YEARS, LAST_10_FINANCIAL_YEARS, LAST_12_MONTHS, 
LAST_3_MONTHS, LAST_6_BIMONTHS, LAST_4_QUARTERS, LAST_2_SIXMONTHS, THIS_FINANCIAL_YEAR,
LAST_FINANCIAL_YEAR, LAST_5_FINANCIAL_YEARS

Custom date periods

Analytics query resources support extra parameters to express periods.

Default pe dimension will fall back to:

• eventDate for /analytics/events/query
• enrollmentDate for /analytics/enrollments/query

Adding conditions on one or more date fields and combining them are allowed.

Usage of custom date periods

In resources supporting custom date periods, there are extra query parameters that will be combined to express conditions on the time dimension.

Conditions can be expressed in the following form:

analytics/events/query/...?...&eventDate=2021&...

It's possible to combine more time fields in the same query:

analytics/events/query/...?...&eventDate=2021&incidentDate=202102&...

All of these conditions can be combined with pe dimension:

analytics/events/query/...?...&dimension=pe:TODAY&enrollmentDate=2021&incidentDate=202102&...

Supported formats are described in "date and period format" above. An extra format is provided to express a range of dates: yyyyMMdd_yyyyMMdd and yyyy-MM-dd_yyyy-MM-dd.

In the example bellow, the endpoint will return events that are scheduled to happen between 20210101 and 20210104:

analytics/events/query/...?...&dimension=pe:TODAY&enrollmentDate=2021&incidentDate=202102&scheduledDate=20210101_20210104&...

Authorities

System authority ids and names can be listed using:

/api/authorities

It returns the following format:

{
  "systemAuthorities": [
    {
      "id": "ALL",
      "name": "ALL"
    },
    {
      "id": "F_ACCEPT_DATA_LOWER_LEVELS",
      "name": "Accept data at lower levels"
    }
  ]
}

Metadata

Identifier schemes

This section provides an explanation of the identifier scheme concept. Identifier schemes are used to map metadata objects to other metadata during import, and to render metadata as part of exports. Note that not all schemes work for all API calls, and not all schemes can be used for both input and output. This is outlined in the sections explaining the various API endpoints.

The full set of identifier scheme object types available are listed below, using the name of the property to use in queries:

• idScheme

• dataElementIdScheme

• categoryOptionComboIdScheme

• orgUnitIdScheme

• programIdScheme

• programStageIdScheme

• trackedEntityIdScheme

• trackedEntityAttributeIdScheme

The general idScheme applies to all types of objects. It can be overridden by specific object types.

The default scheme for all parameters is UID (stable DHIS2 identifiers). The supported identifier schemes are described in the table below.

Note that identifier schemes is not an independent feature but needs to be used in combination with resources such as data value import, metadata import and GeoJson import.

As an example, to specify CODE as the general id scheme and override with UID for organisation unit id scheme you can use these query parameters:

?idScheme=CODE&orgUnitIdScheme=UID

As another example, to specify an attribute for the organisation unit id scheme, code for the data element id scheme and use the default UID id scheme for all other objects you can use these parameters:

?orgUnitIdScheme=ATTRIBUTE:j38fk2dKFsG&dataElementIdScheme=CODE

Browsing the Web API

The entry point for browsing the Web API is /api. This resource provides links to all available resources. Four resource representation formats are consistently available for all resources: HTML, XML, JSON, and JSONP. Some resources will have other formats available, like MS Excel, PDF, CSV, and PNG. To explore the API from a web browser, navigate to the /api entry point and follow the links to your desired resource, for instance /api/dataElements. For all resources which return a list of elements certain query parameters can be used to modify the response:

An example of how these parameters can be used to get a full list of data element groups in XML response format is:

/api/dataElementGroups.xml?links=false&paging=false

You can query for elements on the name property instead of returning a full list of elements using the query query variable. In this example we query for all data elements with the word "anaemia" in the name:

/api/dataElements?query=anaemia

You can get specific pages and page sizes of objects like this:

/api/dataElements.json?page=2&pageSize=20

You can completely disable paging like this:

/api/indicatorGroups.json?paging=false

To order the result based on a specific property:

/api/indicators.json?order=shortName:desc

You can find an object based on its ID across all object types through the identifiableObjects resource:

/api/identifiableObjects/<id>

Translation

DHIS2 supports translations of database content, such as data elements, indicators, and programs. All metadata objects in the Web API have properties meant to be used for display / UI purposes, which include displayName, displayShortName, displayDescription and displayFormName (for data elements and tracked entity attributes).

Translation API

The translations for an object is rendered as part of the object itself in the translations array. Note that the translations array in the JSON/XML payloads is normally pre-filtered for you, which means they can not directly be used to import/export translations (as that would normally overwrite locales other than current users).

Example of data element with translation array filtered on user locale:

{
  "id": "FTRrcoaog83",
  "displayName": "Accute French",
  "translations": [
    {
      "property": "SHORT_NAME",
      "locale": "fr",
      "value": "Accute French"
    },
    {
      "property": "NAME",
      "locale": "fr",
      "value": "Accute French"
    }
  ]
}

Example of data element with translations turned off:

{
  "id": "FTRrcoaog83",
  "displayName": "Accute Flaccid Paralysis (Deaths < 5 yrs)",
  "translations": [
    {
      "property": "FORM_NAME",
      "locale": "en_FK",
      "value": "aa"
    },
    {
      "property": "SHORT_NAME",
      "locale": "en_GB",
      "value": "Accute Flaccid Paral"
    },
    {
      "property": "SHORT_NAME",
      "locale": "fr",
      "value": "Accute French"
    },
    {
      "property": "NAME",
      "locale": "fr",
      "value": "Accute French"
    },
    {
      "property": "NAME",
      "locale": "en_FK",
      "value": "aa"
    },
    {
      "property": "DESCRIPTION",
      "locale": "en_FK",
      "value": "aa"
    }
  ]
}

Note that even if you get the unfiltered result, and are using the appropriate type endpoint i..e /api/dataElements we do not allow updates, as it would be too easy to make mistakes and overwrite the other available locales.

To read and update translations you can use the special translations endpoint for each object resource. These can be accessed by GET or PUT on the appropriate /api/<object-type>/<object-id>/translations endpoint.

As an example, for a data element with identifier FTRrcoaog83, you could use /api/dataElements/FTRrcoaog83/translations to get and update translations. The fields available are property with options NAME, SHORT_NAME, FORM_NAME, DESCRIPTION, locale which supports any valid locale ID and the translated property value.

Example of NAME property for French locale:

{
  "property": "NAME",
  "locale": "fr",
  "value": "Paralysie Flasque Aiguë (Décès <5 ans)"
}

This payload would then be added to a translation array, and sent back to the appropriate endpoint:

{
  "translations": [
    {
      "property": "NAME",
      "locale": "fr",
      "value": "Paralysie Flasque Aiguë (Décès <5 ans)"
    }
  ]
}

For a data element with ID FTRrcoaog83 you can PUT this to /api/dataElements/FTRrcoaog83/translations. Make sure to send all translations for the specific object and not just for a single locale (if not you will potentially overwrite existing locales for other locales).

The status code will be 204 No Content if the data value was successfully saved or updated, or 404 Not Found if there was a validation error (e.g. more than one SHORT_NAME for the same locale).

Web API versions

The Web API is versioned starting from DHIS 2.25. The API versioning follows the DHIS2 major version numbering. As an example, the API version for DHIS 2.33 is 33.

You can access a specific API version by including the version number after the /api component, as an example like this:

/api/33/dataElements

If you omit the version part of the URL, the system will use the current API version. As an example, for DHIS 2.25, when omitting the API part, the system will use API version 25. When developing API clients it is recommended to use explicit API versions (rather than omitting the API version), as this will protect the client from unforeseen API changes.

The last three API versions will be supported. As an example, DHIS version 2.27 will support API version 27, 26 and 25.

Note that the metadata model is not versioned and that you might experience changes e.g. in associations between objects. These changes will be documented in the DHIS2 major version release notes.

Metadata object filter

To filter the metadata there are several filter operations that can be applied to the returned list of metadata. The format of the filter itself is straight-forward and follows the pattern property:operator:value, where property is the property on the metadata you want to filter on, operator is the comparison operator you want to perform and value is the value to check against (not all operators require value).

Please see the schema section to discover which properties are available. In addition to the listed properties filters can apply to custom attribute values by using the attribute's ID as property name.

Recursive filtering, ie. filtering on associated objects or collection of objects, is supported as well.

Operators will be applied as logical and query. If you need a or query, you can have a look at the in filter and the section below. The filtering mechanism allows for recursion. See below for some examples.

Get data elements with id property ID1 or ID2:

/api/dataElements?filter=id:eq:ID1&filter=id:eq:ID2

Get all data elements which have a data set with id ID1:

/api/dataElements?filter=dataSetElements.dataSet.id:eq:ID1

Get all data elements with aggregation operator sum and value type int:

/api/dataElements.json?filter=aggregationOperator:eq:sum&filter=type:eq:int

You can do filtering within collections, e.g. to get data elements which are members of the ANC data element group you can use the following query using the id property of the associated data element groups:

/api/dataElements.json?filter=dataElementGroups.id:eq:qfxEYY9xAl6

To get data elements with a particular attribute value for a metadata attribute, a filter for the attribute ID and the attribute value can be specified using the same collection query syntax:

/api/dataElements.json?filter=attributeValues.attribute.id:eq:n2xYlNbsfko&filter=attributeValues.value:eq:AFP

Get data elements which have any option set:

/api/dataElements?filter=optionSet:!null

Since all operators are and by default, you can't find a data element matching more than one id, for that purpose you can use the in operator.

/api/dataElements.json?filter=id:in:[fbfJHSPpUQD,cYeuwXTCPkU]

Logical operators

As mentioned in the section before, the default logical operator applied to the filters is AND which means that all object filters must be matched. There are however cases where you want to match on one of several filters (maybe id and code field) and in those cases, it is possible to switch the root logical operator from AND to OR using the rootJunction parameter.

Example: Normal filtering where both id and code must match to have a result returned

/api/dataElements.json?filter=id:in:[id1,id2]&filter=code:eq:code1

Example: Filtering where the logical operator has been switched to OR and now only one of the filters must match to have a result returned

/api/dataElements.json?filter=id:in:[id1,id2]&filter=code:eq:code1&rootJunction=OR

Identifiable token filter

In addition to the specific property based filtering mentioned above, we also have token based AND filtering across a set of properties: id, code, and name (also shortName if available). These properties are commonly referred to as identifiable. The idea is to filter metadata whose id, name, code or short name containing something.

Example: Filter all data elements containing 2^nd in any of the following: id,name,code, shortName

/api/dataElements.json?filter=identifiable:token:2nd

It is also possible to specify multiple filtering values.

Example: Get all data elements where ANC visit is found in any of the identifiable properties. The system returns all data elements where both tokens (ANC and visit) are found anywhere in identifiable properties.

/api/dataElements.json?filter=identifiable:token:ANC visit

It is also possible to combine the identifiable filter with property-based filter and expect the rootJunction to be applied.

/api/dataElements.json?filter=identifiable:token:ANC visit&filter=displayName:ilike:tt1

/api/dataElements.json?filter=identifiable:token:ANC visit
  &filter=displayName:ilike:tt1&rootJunction=OR

Indexable only filter for tracked entity attributes

For tracked entity attributes, there is a special filter in addition to the previous mentioned filtering capabilities. Some of the tracked entity attributes are candidates for creating a trigram index for better lookup performance. Using the indexableOnly parameter set to true, the results can be filtered to include only the attributes that are trigram indexable.

Example: Get all tracked entity attributes that are indexable.

/api/trackedEntityAttributtes.json?indexableOnly=true

Additional filters along with the indexableOnly parameter can be specified.

Example: Get all tracked entity attributes where ANC is found in any of the name property. The system returns the tracked entity attributes where the name matches the provided keyword as well as if the attribute is indexable.

/api/trackedEntityAttributtes.json?filter=name:like:ANC&indexableOnly=true

Metadata field filter

In many situations, the default views of the metadata can be too verbose. A client might only need a few fields from each object and want to remove unnecessary fields from the response. To discover which fields are available for each object please see the schema section. In addition to the listed properties custom attributes can be included for top level objects by using the attribute's ID as property name.

The format for include/exclude allows for infinite recursion. To filter at the "root" level you can just use the name of the field, i.e. ?fields=id,name which would only display the id and name fields for every object. For objects that are either collections or complex objects with properties on their own, you can use the format ?fields=id,name,dataSets[id,name] which would return id, name of the root, and the id and name of every data set on that object. Negation can be done with the exclamation operator, and we have a set of presets of field select. Both XML and JSON formats are supported.

Example: Get id and name on the indicators resource:

/api/indicators?fields=id,name

Example: Get id and name from data elements, and id and name from the associated data sets:

/api/dataElements?fields=id,name,dataSets[id,name]

Example: Get id, name and the value of a user defined attribute with ID DnrLSdo4hMl for organisation units:

/api/organisationUnits?fields=id,name,DnrLSdo4hMl

The attribute is then included as property DnrLSdo4hMl of each matching object in the response. This can be renamed using the rename transformer as shown in the next section.

To exclude a field from the output you can use the exclamation ! operator. This is allowed anywhere in the query and will simply not include that property as it might have been inserted in some of the presets.

A few presets (selected fields groups) are available and can be applied using the : operator.

Example: Include all fields from data sets except organisation units:

/api/dataSets?fields=:all,!organisationUnits

Example: Include only id, name and the collection of organisation units from a data set, but exclude the id from organisation units:

/api/dataSets/BfMAe6Itzgt?fields=id,name,organisationUnits[:all,!id]

Example: Include nameable properties from all indicators:

/api/indicators.json?fields=:nameable

Field transformers

Field transforms can be used to transform properties. The syntax is described below.

/api/dataElements/ID?fields=id~rename(i),name~rename(n)

This will rename the id property to i and name property to n.

Multiple transformers can be applied to a single property by repeating the transformer operator:

/api/dataElementGroups.json?fields=id,displayName,dataElements~isNotEmpty~rename(haveDataElements)

The supported transformer operators are described in the table below.

Examples

Examples of transformer usage are found below.

Get the size of a collection:

/api/dataElements?fields=dataSets~size

Test if a collection is empty:

/api/dataElements?fields=dataSets~isEmpty

Test if a collection is not empty:

/api/dataElements?fields=dataSets~isNotEmpty

Rename properties:

/api/dataElements/ID?fields=id~rename(i),name~rename(n)

Apply paging to a collection:

/api/dataElementGroups?fields=id,displayName,dataElements~paging(1;20)

Get array with IDs of organisation units:

/api/categoryOptions.json?fields=id,organisationUnits~pluck

Get array with names of organisation units:

/api/categoryOptions.json?fields=id,organisationUnits~pluck[name]

Key the dataElements array by the id field:

/api/dataElementGroups.json?fields=id,name,dataElements~keyBy[id,name,valueType]

Key the dataElements array by the valueType field, since multiple hits this will results in arrays (of data elements):

/api/dataElementGroups.json?fields=id,name,dataElements~keyBy(valueType)[id,name,valueType]

Metadata create, read, update, delete, validate

All metadata entities in DHIS2 have their own API endpoint which supports CRUD operations (create, read, update and delete). The endpoint URLs follows this format:

/api/<entityName>

The entityName uses the camel-case notation. As an example, the endpoint for data elements is:

/api/dataElements

Create / update parameters

The following request query parameters are available across all metadata endpoints.

Creating and updating objects

For creating new objects you will need to know the endpoint, the type format, and make sure that you have the required authorities. As an example, we will create and update a constant. To figure out the format, we can use the new schema endpoint for getting format description. So we will start with getting that info:

http://<server>/api/schemas/constant.json

From the output, you can see that the required authorities for create are F_CONSTANT_ADD, and the important properties are: name and value. From this, we can create a JSON payload and save it as a file called constant.json:

{
  "name": "PI",
  "value": "3.14159265359"
}

The same content as an XML payload:

<constant name="PI" xmlns="http://dhis2.org/schema/dxf/2.0">
  <value>3.14159265359</value>
</constant>

We are now ready to create the new constant by sending a POST request to the constants endpoint with the JSON payload using curl:

curl -d @constant.json "http://server/api/constants" -X POST
  -H "Content-Type: application/json" -u user:password

A specific example of posting the constant to the demo server:

curl -d @constant.json "https://play.dhis2.org/api/constants" -X POST
  -H "Content-Type: application/json" -u admin:district

If everything went well, you should see an output similar to:

{
  "status": "SUCCESS",
  "importCount": {
    "imported": 1,
    "updated": 0,
    "ignored": 0,
    "deleted": 0
  },
  "type": "Constant"
}

The process will be exactly the same for updating, you make your changes to the JSON/XML payload, find out the ID of the constant, and then send a PUT request to the endpoint including ID:

curl -X PUT -d @pi.json -H "Content-Type: application/json"
  -u user:password "http://server/api/constants/ID"

Deleting objects

Deleting objects is very straight forward, you will need to know the ID and the endpoint of the type you want to delete, let's continue our example from the last section and use a constant. Let's assume that the id is abc123, then all you need to do is the send the DELETE request to the endpoint + id:

curl -X DELETE -u user:password "http://server/api/constants/ID"

A successful delete should return HTTP status 204 (no content).

Adding and removing objects in collections

The collections resource lets you modify collections of objects.

Adding or removing single objects

In order to add or remove objects to or from a collection of objects you can use the following pattern:

/api/{collection-object}/{collection-object-id}/{collection-name}/{object-id}

You should use the POST method to add, and the DELETE method to remove an object. When there is a many-to-many relationship between objects, you must first determine which object owns the relationship. If it isn't clear which object this is, try the call both ways to see which works.

The components of the pattern are:

• collection object: The type of objects that owns the collection you want to modify.

• collection object id: The identifier of the object that owns the collection you want to modify.

• collection name: The name of the collection you want to modify.

• object id: The identifier of the object you want to add or remove from the collection.

As an example, in order to remove a data element with identifier IDB from a data element group with identifier IDA you can do a DELETE request:

DELETE /api/dataElementGroups/IDA/dataElements/IDB

To add a category option with identifier IDB to a category with identifier IDA you can do a POST request:

POST /api/categories/IDA/categoryOptions/IDB

Adding or removing multiple objects

You can add or remove multiple objects from a collection in one request with a payload like this:

{
  "identifiableObjects": [{
      "id": "IDA"
    }, {
      "id": "IDB"
    }, {
      "id": "IDC"
    }
  ]
}

Using this payload you can add, replace or delete items:

Adding Items:

POST /api/categories/IDA/categoryOptions

Replacing Items:

PUT /api/categories/IDA/categoryOptions

Delete Items:

DELETE /api/categories/IDA/categoryOptions

Adding and removing objects in a single request

You can both add and remove objects from a collection in a single POST request to the following URL:

POST /api/categories/IDA/categoryOptions

The payload format is:

{
  "additions": [{
      "id": "IDA"
    }, {
      "id": "IDB"
    }, {
      "id": "IDC"
    }
  ],
  "deletions": [{
      "id": "IDD"
    }, {
      "id": "IDE"
    }, {
      "id": "IDF"
    }
  ]
}

Validating payloads

DHIS2 supports system wide validation of metadata payloads, which means that create and update operations on the API endpoints will be checked for valid payload before allowing changes to be made. To find out what validations are in place for a specific endpoint, have a look at the /api/schemas endpoint, i.e. to figure out which constraints a data element have, you would go to /api/schemas/dataElement.

You can also validate your payload manually by sending it to the proper schema endpoint. If you wanted to validate the constant from the create section before, you would send it like this:

POST /api/schemas/constant

A simple (non-validating) example would be:

curl -X POST -d "{\"name\": \"some name\"}" -H "Content-Type: application/json"
  -u admin:district "https://play.dhis2.org/dev/api/schemas/dataElement"

Which will yield the result:

[
   {
      "message" : "Required property missing.",
      "property" : "type"
   },
   {
      "property" : "aggregationOperator",
      "message" : "Required property missing."
   },
   {
      "property" : "domainType",
      "message" : "Required property missing."
   },
   {
      "property" : "shortName",
      "message" : "Required property missing."
   }
]

Partial updates

For our API endpoints that deal with metadata, we support partial updates (PATCH) using the JSON patch standard. The payload basically outlines a set of operation you want applied to a existing metadata object. For JSON patch details and examples, see jsonpatch.com. Three operators are supported: add, remove and replace.

Below is a few examples relevant to DHIS2. Note that any update to a payload should be thought of as a HTTP PUT operation, i.e. any mutation must result in a valid PUT metadata payload.

The default importReportMode for JSON patch is ERRORS_NOT_OWNER which implies that when updating any property which is not owned by that particular object (for example trying to add a indicator group directly to an indicator) you will get an error.

As per the JSON patch specification you must always use the mimetype application/json-patch+json when sending patches.

Examples

Update name and value type of data element

PATCH /api/dataElements/{id}

[
  {"op": "add", "path": "/name", "value": "New Name"},
  {"op": "add", "path": "/valueType", "value": "INTEGER"}
] 

Add new data element to a data element group

PATCH /api/dataElementGroups/{id}

[
  {"op": "add", "path": "/dataElements/-", "value": {"id": "data-element-id"}}
]

Remove all data element associations from a data element group

PATCH /api/dataElementGroups/{id}

[
  {"op": "remove", "path": "/dataElements"}
]

Change domain and value type of a data element

PATCH /api/dataElements/{id}

[
    {"op": "add", "path": "/domainType", "value": "TRACKER"},
    {"op": "add", "path": "/valueType", "value": "INTEGER"}
]

Remove a specific orgUnit from an orgUnit group

PATCH /api/organisationUnitGroups/{id}

[
  {"op": "remove", "path": "/organisationUnits/1"}
]

Blocked add dataElementGroup to dataElement

PATCH /api/dataElements/{id}?importReportMode=ERRORS_NOT_OWNER

[
    {"op": "add", "path": "/dataElementGroups/-", "value": {"id": "data-element-group-id"}}
]

Blocked update name of dataElementGroup in dataElement

PATCH /api/dataElements/{id}?importReportMode=ERRORS_NOT_OWNER

[
    {"op": "add", "path": "/dataElementGroups/0", "value": {"name": "new-name"}}
]

Remove collection item by id

PATCH /api/dataSets/{id}?importReportMode=ERRORS_NOT_OWNER

[
    {"op": "remove-by-id", "path": "/organisationUnits", "id": "u6CvKyF0Db5"}
]

Patch request with invalid path

If path property is invalid or does not exist the patch service will return an error as below

PATCH /api/dataSets/{id}?importReportMode=ERRORS_NOT_OWNER

[
    {"op": "remove-by-id", "path": "/test", "id": "u6CvKyF0Db5"}
]

{
    "httpStatus": "Bad Request",
    "httpStatusCode": 400,
    "status": "ERROR",
    "message": "Invalid path /test"
}

Metadata CSV export

Field filtering works almost the same for CSV (please note that using CSV on the /api/metadata endpoint is not supported), but not that field fransformations are not yet supported.

For endpoints that support CSV (our metadata endpoints like /api/dataElements /api/organisationUnits) you can either use the Accept header with value text/csv or you can use the extension .csv. Be aware that complex objects are not supported, and we only support id-object collections (so a list of UIDs will be returned).

Examples

Get all data elements including their group associations

/api/dataElements.csv?fields=id,displayName,dataElementGroups

Get all org units including geometry (which will get ignored)

/api/organisationUnits.csv?fields=id,displayName,organisationUnitGroups,geometry

Metadata export

This section explains the metatada API which is available at /api/metadata. XML and JSON resource representations are supported.

/api/metadata

The most common parameters are described below in the "Export Parameter" table. You can also apply this to all available types by using type:fields=<filter> and type:filter=<filter>. You can also enable/disable the export of certain types by setting type=true|false.

Metadata export examples

Export all metadata. Be careful as the response might be very large depending on your metadata configuration:

/api/metadata

Export all metadata ordered by lastUpdated descending:

/api/metadata?defaultOrder=lastUpdated:desc

Export metadata only including indicators and indicator groups:

/api/metadata?indicators=true&indicatorGroups=true

Export id and displayName for all data elements, ordered by displayName:

/api/metadata?dataElements:fields=id,name&dataElements:order=displayName:desc

Export data elements and indicators where name starts with "ANC":

/api/metadata?filter=name:startsWith:ANC&dataElements=true&indicators=true

Metadata export with dependencies

When you want to exchange metadata for a data set, program, category combo, dashboard, option set or data element group from one DHIS2 instance to another instance there are six dedicated endpoints available:

/api/dataSets/{id}/metadata.json

/api/programs/{id}/metadata.json

/api/categoryCombos/{id}/metadata.json

/api/dashboards/{id}/metadata.json

/api/optionSets/{id}/metadata.json

/api/dataElementGroups/{id}/metadata.json

These exports can then be imported using /api/metadata.

These endpoints also support the following parameters:

Metadata import

This section explains the metadata import API. XML and JSON resource representations are supported. Metadata can be imported using a POST request.

/api/metadata

The importer allows you to import metadata payloads which may include many different entities and any number of objects per entity. The metadata export generated by the metadata export API can be imported directly.

The metadata import endpoint support a variety of parameters, which are listed below.

(*) Currently the mergeMode=MERGE option of the import service has limitations and doesn't support all objects. It doesn't work with some object types such as Embedded objects, or objects which are saved as JSONB format in database ( sharing, attributeValues, etc...). Fixing those issues are complicated and would just cause new issues. Therefore, this mergedMode=MERGE is deprecated and currently is not recommended to use. The update mode should always be mergedMode=REPLACE. We have developed a new JSON Patch API which can be used as an alternative approach. This feature is introduced in 2.37 release.

An example of a metadata payload to be imported looks like this. Note how each entity type have their own property with an array of objects:

{
  "dataElements": [
    {
      "name": "EPI - IPV 3 doses given",
      "shortName": "EPI - IPV 3 doses given",
      "aggregationType": "SUM",
      "domainType": "AGGREGATE",
      "valueType": "INTEGER_ZERO_OR_POSITIVE"
    },
    {
      "name": "EPI - IPV 4 doses given",
      "shortName": "EPI - IPV 4 doses given",
      "aggregationType": "SUM",
      "domainType": "AGGREGATE",
      "valueType": "INTEGER_ZERO_OR_POSITIVE"
    }
  ],
  "indicators": [
    {
      "name": "EPI - ADS stock used",
      "shortName": "ADS stock used",
      "numerator": "#{LTb8XeeqeqI}+#{Fs28ZQJET6V}-#{A3mHIZd2tPg}",
      "numeratorDescription": "ADS 0.05 ml used",
      "denominator": "1",
      "denominatorDescription": "1",
      "annualized": false,
      "indicatorType": {
        "id": "kHy61PbChXr"
      }
    }
  ]
}

When posting this payload to the metadata endpoint, the response will contain information about the parameters used during the import and a summary per entity type including how many objects were created, updated, deleted and ignored:

{
  "importParams": {
    "userOverrideMode": "NONE",
    "importMode": "COMMIT",
    "identifier": "UID",
    "preheatMode": "REFERENCE",
    "importStrategy": "CREATE_AND_UPDATE",
    "atomicMode": "ALL",
    "mergeMode": "REPLACE",
    "flushMode": "AUTO",
    "skipSharing": false,
    "skipTranslation": false,
    "skipValidation": false,
    "metadataSyncImport": false,
    "firstRowIsHeader": true,
    "username": "UNICEF_admin"
  },
  "status": "OK",
  "typeReports": [
    {
      "klass": "org.hisp.dhis.dataelement.DataElement",
      "stats": {
        "created": 2,
        "updated": 0,
        "deleted": 0,
        "ignored": 0,
        "total": 2
      }
    },
    {
      "klass": "org.hisp.dhis.indicator.Indicator",
      "stats": {
        "created": 1,
        "updated": 0,
        "deleted": 0,
        "ignored": 0,
        "total": 1
      }
    }
  ],
  "stats": {
    "created": 3,
    "updated": 0,
    "deleted": 0,
    "ignored": 0,
    "total": 3
  }
}

GeoJSON import

The GeoJSON import is used to attach geometry data to organisation units.

For a bulk import a GeoJSON file with a feature collection is expected. Each feature in the collection requires a reference to the organisation unit it should be linked to.

By default, the geometry from the file is stored as the geometry property of an organisation unit. To store additional geometries attributes of type GEOJSON can be created. When attributes are use all geometries from a file are stored for the same attribute which is provided with an additional parameter attributeId.

GeoJSON Bulk Data Import

Uasge:

POST /api/organisationUnits/geometry

The post body is the GeoJSON file. Content type should be application/json or application/geo+json. The file may be .zip or .gzip compressed.

For example, a default file where id is used to refer to an organisation unit id has this structure:

{ 
  "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "id": "O6uvpzGd5pu",
      "geometry": { ... }
    },
    ...
  ]
}

A file where a feature property is used to refer to the organisation unit code would have this structure:

{ 
  "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "properties": { "code": "OU1_CODE" },
      "geometry": { ... }
    },
    ...
  ]
}

A geometry may also be null to effectively clear or delete the geometry for specific organisation units. There is a special bulk deletion API that is described in the next section.

When run synchronously an import report is returned directly. The HTTP status code is always OK, the status in the message payload indicates if all rows were imported successfully. The import counts statistics contained in the report give further information:

• imported: number of organisation units that were successfully updated with a geometry that did not have one before for the updated property
• updated: number of organisation units that were successfully updated with a geometry that did have value for the updated property already
• ignored: number of organisation units that failed to update
• deleted: number of organisation units that where successfully update with a empty geometry

When the import is run asynchronous the request returns immediately with status OK and job configuration response that contains a relative reference to the task endpoint that allows to track the status of the asynchronous import. For example:

/api/system/tasks/GEOJSON_IMPORT/{job-id}

The summary that is returned directly for synchronous execution is available at

/api/system/taskSummaries/GEOJSON_IMPORT/{job-id}

once the import is finished.

GeoJSON Bulk Data Deletion

To clear or unset the geometry data for all organisation units use:

DELETE /api/organisationUnits/geometry

To clear or unset the geometry data for a specific GEOJSON attribute for all organisation units use:

DELETE /api/organisationUnits/geometry?attributeId={attr-id}

Clearing is always synchronous and returns a similar report as the bulk import. It does not support any other parameters. No dry-run can be performed. Bulk clearing requires the F_PERFORM_MAINTENANCE authority.

GeoJSON Single Data Import

The single import allows to update the geometry of a single organisation unit.

POST /api/organisationUnits/{id}/geometry

The post body only contains the GeoJSON geometry value, for example:

{
  "type": "Polygon",
  "coordinates": [...]
}

GeoJSON Single Data Deletion

To clear the geometry GeoJSON data of an individual organisation unit use:

DELETE /api/organisationUnits/{id}/geometry

Similarly to clear a GEOJSON attribute value for an individual organisation unit use:

DELETE /api/organisationUnits/{id}/geometry?attributeId={attr-id}

Clearing is always synchronous returns a similar report as single import. The dry-run parameter is supported as well. The performing user requires authority to modify the target organisation unit.

Schema

A resource which can be used to introspect all available DXF 2 objects can be found on /api/schemas. For specific resources you can have a look at /api/schemas/<type>.

To get all available schemas in XML:

GET /api/schemas.xml

To get all available schemas in JSON:

GET /api/schemas.json

To get JSON schema for a specific class:

GET /api/schemas/dataElement.json

Icons

DHIS2 includes a collection of icons that can be used to give visual context to metadata. These icons can be accessed through the icons resource.

GET /api/icons

This endpoint returns a list of information about the available icons. Each entry contains information about the icon, and a reference to the actual icon.

{
  key: "mosquito_outline",
  description: "Mosquito outline",
  keywords: [
    "malaria",
    "mosquito",
    "dengue"
  ],
  href: "<dhis server>/api/icons/mosquito_outline/icon.svg"
}

The keywords can be used to filter which icons to return. Passing a list of keywords with the request will only return icons that match all the keywords:

GET /api/icons?keywords=shape,small

A list of all unique keywords can be found at the keywords resource:

GET /api/icons/keywords

Render type

Some metadata types have a property named renderType. The render type property is a map between a device and a renderingType. Applications can use this information as a hint on how the object should be rendered on a specific device. For example, a mobile device might want to render a data element differently than a desktop computer.

There is currently two different kinds of renderingTypes available:

1. Value type rendering

2. Program stage section rendering

There is also 2 device types available:

1. MOBILE

2. DESKTOP

The following table lists the metadata and rendering types available. The value type rendering has addition constraints based on the metadata configuration, which will be shown in a second table.

Since handling the default rendering of data elements and tracked entity attributes are depending on the value type of the object, there is also a DEFAULT type to tell the client it should be handled as normal. Program Stage Section is LISTING as default.

A complete reference of the previous table can also be retrieved using the following endpoint:

GET /api/staticConfiguration/renderingOptions

Value type rendering also has some additional properties that can be set, which is usually needed when rendering some of the specific types:

The renderingType can be set when creating or updating the metadata listed in the first table. An example payload for the rendering type for program stage section looks like this:

{
  "renderingType": {
    "type": "MATRIX"
  }
}

For data element and tracked entity attribute:

{
  "renderingType": {
    "type": "SLIDER",
    "min": 0,
    "max": 1000,
    "step": 50,
    "decimalPoints": 0
  }
}

Object Style

Most metadata have a property names "style". This property can be used by clients to represent the object in a certain way. The properties currently supported by style is as follows:

Currently, there is no official list or support for icon-libraries, so this is currently up to the client to provide. The following list shows all objects that support style:

• Data element

• Data element category option

• Data set

• Indicator

• Option

• Program

• Program Indicator

• Program Section

• Program Stage

• Program Stage Section

• Relationship (Tracker)

• Tracked Entity Attribute

• Tracked Entity Type

When creating or updating any of these objects, you can include the following payload to change the style:

{
  "style": {
    "color": "#ffffff",
    "icon": "my-beautiful-icon"
  }
}

Indicators

This section describes indicators and indicator expressions.

Aggregate indicators

To retrieve indicators you can make a GET request to the indicators resource like this:

/api/indicators

Indicators represent expressions which can be calculated and presented as a result. The indicator expressions are split into a numerator and denominator. The numerators and denominators are mathematical expressions which can contain references to data elements, other indicators, constants and organisation unit groups. The variables will be substituted with data values when used e.g. in reports. Variables which are allowed in expressions are described in the following table.

Within a Data element operand or an Aggregate data element, the following substitutions may be made:

The syntax looks like this:

#{<dataelement-id>.<catoptcombo-id>} + C{<constant-id>} + OUG{<orgunitgroup-id>}

A corresponding example looks like this:

#{P3jJH5Tu5VC.S34ULMcHMca} + C{Gfd3ppDfq8E} + OUG{CXw2yu5fodb}

Note that for data element variables the category option combo identifier can be omitted. The variable will then represent the total for the data element, e.g. across all category option combos. Example:

#{P3jJH5Tu5VC} + 2

Data element operands can include any of category option combination and attribute option combination, and use wildcards to indicate any value:

#{P3jJH5Tu5VC.S34ULMcHMca} + #{P3jJH5Tu5VC.*.j8vBiBqGf6O} + #{P3jJH5Tu5VC.S34ULMcHMca.*}

An example using a data element group:

#{deGroup:oDkJh5Ddh7d} + #{deGroup:GBHN1a1Jddh.j8vBiBqGf6O}

An example using a category option, data element group, and a category option group:

#{P3jJH5Tu5VC.co:FbLZS3ueWbQ} + #{deGroup:GBHN1a1Jddh.coGroup:OK2Nr4wdfrZ.j8vBiBqGf6O}

An example using multiple category option groups:

#{P3jJH5Tu5VC.coGroup:OK2Nr4wdfrZ&j3C417uW6J7&ddAo6zmIHOk}

An example using a program data element and a program attribute:

( D{eBAyeGv0exc.vV9UWAZohSf} * A{IpHINAT79UW.cejWyOfXge6} ) / D{eBAyeGv0exc.GieVkTxp4HH}

An example combining program indicators and aggregate indicators:

I{EMOt6Fwhs1n} * 1000 / #{WUg3MYWQ7pt}

An example using a reporting rate:

R{BfMAe6Itzgt.REPORTING_RATE} * #{P3jJH5Tu5VC.S34ULMcHMca}

Another reporting rate example using actual data set reports and expected reports:

R{BfMAe6Itzgt.ACTUAL_REPORTS} / R{BfMAe6Itzgt.EXPECTED_REPORTS}

An example using an existing indicator:

N{Rigf2d2Zbjp} * #{P3jJH5Tu5VC.S34ULMcHMca}

Expressions can be any kind of valid mathematical expression, as an example:

( 2 * #{P3jJH5Tu5VC.S34ULMcHMca} ) / ( #{FQ2o8UBlcrS.S34ULMcHMca} - 200 ) * 25

Program indicators

To retrieve program indicators you can make a GET request to the program indicators resource like this:

/api/programIndicators

Program indicators can contain information collected in a program. Indicators have an expression which can contain references to data elements, attributes, constants and program variables. Variables which are allowed in expressions are described in the following table.

The syntax looks like this:

#{<programstage-id>.<dataelement-id>} + #{<attribute-id>} + V{<varible-id>} + C{<constant-id>}

A corresponding example looks like this:

#{A03MvHHogjR.a3kGcGDCuk6} + A{OvY4VVhSDeJ} + V{incident_date} + C{bCqvfPR02Im}

Expressions

Expressions are mathematical formulas which can contain references to data elements, constants and organisation unit groups. To validate and get the textual description of an expression, you can make a GET request to the expressions resource:

/api/expressions/description?expression=<expression-string>

The response follows the standard JSON web message format. The status property indicates the outcome of the validation and will be "OK" if successful and "ERROR" if failed. The message property will be "Valid" if successful and provide a textual description of the reason why the validation failed if not. The description provides a textual description of the expression.

{
  "httpStatus": "OK",
  "httpStatusCode": 200,
  "status": "OK",
  "message": "Valid",
  "description": "Acute Flaccid Paralysis"
}

Organisation units

The organisationUnits resource follows the standard conventions as other metadata resources in DHIS2. This resource supports some additional query parameters.

Get list of organisation units

To get a list of organisation units you can use the following resource.

/api/33/organisationUnits

Get organisation unit with sub-hierarchy

To get an organisation unit including organisation units in its sub-hierarchy you can use the following resource.

/api/33/organisationUnits/{id}

Get organisation units by category option

Purpose-built endpoint to retrieve associations between category options and organisation units. This endpoint is the preferred way to retrieve program organisation unit associations.

/api/33/categoryOptions/orgUnits?categoryOptions={categoryOptionIdA},{categoryOptionIdB}

responses will have the following format:

{
  "<categoryOptionIdA>": [
    "<orgUnitUid>",
    "<orgUnitUid>"
  ],
  "<categoryOptionIdB>": [
    "<orgUnitUid>",
    "<orgUnitUid>"
  ],
  "<categoryOptionIdC>": []
}

Category options that are accessible by all organisation units are returned with an empty array ([]) of organisation units.

Get organisation units by programs

Purpose-built endpoint to retrieve associations between programs and organisation units. This endpoint is the preferred way to retrieve program organisation unit associations.

/api/33/programs/orgUnits?programs={programIdA},{programIdB}

responses will have the following format:

{
  "<programIdA>": [
    "<orgUnitUid>",
    "<orgUnitUid>"
  ],
  "<programIdB>": [
    "<orgUnitUid>",
    "<orgUnitUid>"
  ],
  "<programIdC>": []
}

Programs which are accessible by all organisation units are returned with an empty array ([]) of organisation units.

Split organisation unit

The organisation unit split endpoint allows you to split organisation units into a number of target organisation units.

Request

Split organisation units with a POST request:

POST /api/organisationUnits/split

The payload in JSON format looks like the following:

{
  "source": "rspjJHg4WY1",
  "targets": [
    "HT0w9YLMLyn",
    "rEpnzuNpRKM"
  ],
  "primaryTarget": "HT0w9YLMLyn",
  "deleteSource": true
}

The JSON properties are described in the following table.

The split operation will split the source org unit into the target org units. It is recommended to first create new target org units before performing the split, and at a minimum ensure that no aggregate data exists for the target org units. Any number of target org units can be specified.

The split operation will transfer all of the metadata associations of the source org unit over to the target org units. This includes data sets, programs, org unit groups, category options, users, visualizations, maps and event reports.

The operation will transfer all data records of the source org unit over to the org unit specified as the primary target, or if not specified, the first specified target org unit. This includes aggregate data values, data approval records, events, tracked entities and more.

Validation

The following constraints and error codes apply.

Merge organisation units

The organisation unit merge endpoint allows you to merge a number of organisation units into a target organisation unit.

Authorisation

The main authority required to perform an organisation unit merge is F_ORGANISATION_UNIT_MERGE.

Request

Merge organisation units with a POST request:

POST /api/organisationUnits/merge

The payload in JSON format looks like the following:

{
  "sources": [
    "jNb63DIHuwU",
    "WAjjFMDJKcx"
  ],
  "target": "V9rfpjwHbYg",
  "dataValueMergeStrategy": "LAST_UPDATED",
  "dataApprovalMergeStrategy": "LAST_UPDATED",
  "deleteSources": true
}

The JSON properties are described in the following table.

The merge operation will merge the source org units into the target org unit. It is recommended to first create a new target org unit before performing the merge, and at a minimum ensure that no aggregate data exists for the target org unit. Any number of source org units can be specified.

The merge operation will transfer all of the metadata associations of the source org units over to the target org unit. This includes data sets, programs, org unit groups, category options, users, visualizations, maps and event reports. The operation will also transfer all event and tracker data, such as events, enrollments, ownership history, program ownership and tracked entities, over to the target org unit.

The specified data value merge strategy defines how data values are handled. For strategy LAST_UPDATED, data values for all source org units are transferred over to the target org unit, and in situation where data values exist for the same parameters, the last updated or created data value will be used. This is done to avoid duplication of data. For strategy DISCARD, data values are not transferred over to the target org unit, and simply deleted. The specified data approval merge strategy defines how data approval records are handled, and follows the same logic as data values.

Validation

The following constraints and error codes apply.

Data sets

The dataSets resource follows the standard conventions as other metadata resources in DHIS2. This resource supports some additional query parameters.

/api/33/dataSets

To retrieve the version of a data set you can issue a GET request:

GET /api/33/dataSets/<uid>/version

To bump (increase by one) the version of a data set you can issue a POST request:

POST /api/33/dataSets/<uid>/version

Data set notification template

The dataset notification templates resource follows the standard conventions as other metadata resources in DHIS2.

GET /api/33/dataSetNotficationTemplates

To retrieve data set notification template you can issue a GET request:

GET /api/33/dataSetNotficationTemplates/<uid>

To add data set notification template you can issue a POST request:

POST /api/33/dataSetNotficationTemplates

To delete data set notification template you can issue a DELETE request:

DELETE /api/33/dataSetNotficationTemplates/<uid>

JSON payload sample is given below:

{
  "name": "dataSetNotificationTemplate1",
  "dataSetNotificationTrigger": "DATA_SET_COMPLETION",
  "relativeScheduledDays": 0,
  "notificationRecipient": "ORGANISATION_UNIT_CONTACT",
  "dataSets": [{
    "id": "eZDhcZi6FLP"
  }],
  "deliveryChannels": ["SMS","EMAIL"],
  "subjectTemplate": "V{data_set_name}",
  "messageTemplate": "V{data_set_name}V{registration_period}",
  "sendStrategy": "SINGLE_NOTIFICATION"
}

notificationRecipient can be one of: - USER_GROUP for internal messages - ORGANISATION_UNIT_CONTACT for external messages

Filled organisation unit levels

The filledOrganisationUnitLevels resource provides an ordered list of organisation unit levels, where generated levels are injected into the list to fill positions for which it does not exist a persisted level.

GET /api/33/filledOrganisationUnitLevels

To set the organisation unit levels you can issue a POST request with a JSON payload and content type application/json looking like this:

{
  "organisationUnitLevels": [{
    "name": "National",
    "level": 1,
    "offlineLevels": 3
  }, {
    "name": "District",
    "level": 2
  }, {
    "name": "Chiefdom",
    "level": 3
  }, {
    "name": "Facility",
    "level": 4
  }]
}

Predictors

A predictor allows you to generate data values based on an expression. This can be used for example to generate targets, thresholds, or estimated values.

To retrieve predictors you can make a GET request to the predictors resource like this:

/api/predictors

Creating a predictor

You can create a predictor with a POST request to the predictors resource:

POST /api/predictors

A sample payload looks like this:

{
  "id": "AG10KUJCrRk",
  "name": "Malaria Outbreak Threshold Predictor",
  "shortName": "Malaria Outbreak Predictor",
  "description": "Computes the threshold for potential malaria outbreaks based on the mean plus 1.5x the std dev",
  "output": {
    "id": "nXJJZNVAy0Y"
  },
  "generator": {
    "expression": "AVG(#{r6nrJANOqMw})+1.5*STDDEV(#{r6nrJANOqMw})",
    "description": "Maximum normal malaria case count",
    "missingValueStrategy": "NEVER_SKIP",
    "slidingWindow": false
  },
  "periodType": "Monthly",
  "sequentialSampleCount": 4,
  "sequentialSkipCount": 1,
  "annualSampleCount": 3,
  "organisationUnitLevels": [4]
}

The output element refers to the identifier of the data element for which to saved predicted data values. The generator element refers to the expression to use when calculating the predicted values.

Predictor expressions

A predictor always has a generator expression that describes how the predicted value is calculated. A predictor may also have a skip test expression returning a boolean value. When the skip test expression is present, it is evaluated in each of the sampled periods to tell whether values from that period should be skipped.

The following variables may be used in either a generator expression or a skip test expression:

Generating predicted values

To run all predictors (generating predicted values) you can make a POST request to the run resource:

POST /api/predictors/run

To run a single predictor you can make a POST request to the run resource for a predictor:

POST /api/predictors/AG10KUJCrRk/run

Program rules

This section is about sending and reading program rules, and explains the program rules data model. The program rules give functionality to configure dynamic behaviour in the programs in DHIS2.

Program rule model

The program rules data model consists of programRuleVariables, programRules and programRuleActions. The programRule contains an expression - when this expression is true, the child programRuleActions is triggered. The programRuleVariables is used to address data elements, tracked entity data values and other data values needed to run the expressions. All programRules in a program share the same library of programRuleVariables, and one programRuleVariable can be used in several programRules' expressions.

Program rule model details

The following table gives a detailed overview over the programRule model.

Program rule action model details

The following table gives a detailed overview over the programRuleAction model.

ProgramRuleAction Validation

There are certain validations added to ProgramRuleAction model in 2.37. Main purpose was to keep user from creating erroneous ProgramRules in order to keep the database consistent. These validations depends on program rule action type. Each action type has its own respective validation.

Apart from above validations, data field in program rule action which normally contains expression can also be evaluated using below api endpoint.

POST /api/programRuleActions/data/expression/description?programId=<uid>

{
  "condition": "1 + 1"
}

Program rule variable model details

The following table gives a detailed overview over the programRuleVariable model.

Creating program rules

• To perform crud operations, programRules resource is available in API.

To retrieve list of programRules you can do a GET request like this:

/api/programRules

To retrieve single programRule you can do a GET request like this:

/api/programRules/<program_rule_uid>

To save/add single programRule you can do a POST request like this:

/api/programRules/<program_rule_uid>

To update single programRule you can do a PUT request like this:

/api/programRules/<program_rule_uid>

To delete single programRule you can do a DELETE request like this:

/api/programRules/<program_rule_uid>

To retrieve description of programRule condition you can use POST and provide condition string in the POST body.

/api/programRules/condition/description?<program_rule_uid>

Forms

To retrieve information about a form (which corresponds to a data set and its sections) you can interact with the form resource. The form response is accessible as XML and JSON and will provide information about each section (group) in the form as well as each field in the sections, including labels and identifiers. By supplying period and organisation unit identifiers the form response will be populated with data values.

To retrieve the form for a data set you can do a GET request like this:

/api/dataSets/<dataset-id>/form.json

To retrieve the form for the data set with identifier "BfMAe6Itzgt" in XML:

/api/dataSets/BfMAe6Itzgt/form

To retrieve the form including metadata in JSON:

/api/dataSets/BfMAe6Itzgt/form.json?metaData=true

To retrieve the form filled with data values for a specific period and organisation unit in XML:

/api/dataSets/BfMAe6Itzgt/form.xml?ou=DiszpKrYNg8&pe=201401

When it comes to custom data entry forms, this resource also allows for creating such forms directly for a data set. This can be done through a POST or PUT request with content type text/html where the payload is the custom form markup such as:

curl -d @form.html "localhost/api/dataSets/BfMAe6Itzgt/form"
  -H "Content-Type:text/html" -u admin:district -X PUT

Documents

References to files can be stored with the document resource.

A GET request to the documents endpoint will return all documents:

/api/documents

A POST request to the documents endpoint will create a new document:

curl -X POST -d @document.json -H "Content-type: application/json"
  "http://dhis.domain/api/documents"

{
  "name": "dhis home",
  "external": true,
  "url": "https://www.dhis2.org"
}

A GET request with the id of a document appended will return information about the document. A PUT request to the same endpoint will update the fields of the document:

/api/documents/<documentId>

Appending /data to the GET request will return the actual file content of the document:

/api/documents/<documentId>/data

CSV metadata import

DHIS2 supports import of metadata in the CSV format, such as data elements, organisation units and validation rules. Properties for the various metadata objects are identified based on the column order/column index (see below for details). You can omit non-required object properties/columns, but since the column order is significant, an empty column must be included. In other words, if you would like to specify properties/columns which appear late in the column order but not specify certain columns which appear early in the order you can include empty/blank columns for them.

The first row of the CSV file is considered to be a header and is ignored during import. The comma character should be used as a text delimiter. Text which contains commas must be enclosed in double quotes.

To upload metadata in CSV format you can make a POST request to the metadata endpoint:

POST /api/metadata?classKey=CLASS-KEY

The following object types are supported. The classKey query parameter is mandatory and can be found next to each object type in the table below.

Tip

If using curl, the --data-binary option should be used as it preserves line breaks and newlines, which is essential for CSV data.

As an example, to upload a file of data elements in CSV format with curl you can use the following command:

curl --data-binary @data_elements.csv "http://localhost/api/metadata?classKey=DATA_ELEMENT"
  -H "Content-Type:application/csv" -u admin:district

The formats for the currently supported object types for CSV import are listed in the following sections.

Data elements

An example of a CSV file for data elements can be seen below. The first row will always be ignored. Note how you can skip columns and rely on default values to be used by the system. You can also skip columns which you do not use which appear to the right of the ones

name,uid,code,shortname,description
"Women participated skill development training",,"D0001","Women participated in training"
"Women participated community organizations",,"D0002","Women participated in organizations"

Organisation units

A minimal example for importing organisation units with a parent unit looks like this:

name,uid,code,parent
"West province",,"WESTP","ImspTQPwCqd"
"East province",,"EASTP","ImspTQPwCqd"

Validation rules

Option sets

The format for option sets is special. The three first values represent an option set. The three last values represent an option. The first three values representing the option set should be repeated for each option.

optionsetname,optionsetuid,optionsetcode,optionname,optionuid,optioncode
"Color",,"COLOR","Blue",,"BLUE"
"Color",,"COLOR","Green",,"GREEN"
"Color",,"COLOR","Yellow",,"YELLOW"
"Sex",,,"Male",,"MALE"
"Sex",,,"Female",,"FEMALE"
"Sex",,,"Unknown",,"UNKNOWN"
"Result",,,"High",,"HIGH"
"Result",,,"Medium",,"MEDIUM"
"Result",,,"Low",,"LOW"
"Impact","cJ82jd8sd32","IMPACT","Great",,"GREAT"
"Impact","cJ82jd8sd32","IMPACT","Medium",,"MEDIUM"
"Impact","cJ82jd8sd32","IMPACT","Poor",,"POOR"

Option group

Sample OptionGroup CSV payload

optionGroupName,optionGroupUid,optionGroupCode,optionGroupShortName,optionSetUid,optionUid,optionCode
optionGroupA,,,groupA,xmRubJIhmaK,,OptionA
optionGroupA,,,groupA,xmRubJIhmaK,,OptionB
optionGroupB,,,groupB,QYDAByFgTr1,,OptionC

Option Group Set

Sample OptionGroupSet CSV payload

name,uid,code,description,datadimension,optionsetuid
optiongroupsetA,,,,,xmRubJIhmaK
optiongroupsetB,,,,false,QYDAByFgTr1

Indicators

An example of a CSV file for Indicators can be seen below. The first row will always be ignored. Note how you can skip columns and rely on default values to be used by the system. You can also skip columns which you do not use which appear to the right of the ones

Name,UID,Code,Description,shortName,denominator,denominatorDescription,numerator,numeratorDescription,annualized,decimals,indicatorType
Indicator A,yiAKjiZVoOU,CodeA,Indicator A description,Indicator A shortname,#{fbfJHSPpUQD},denominatorDescription,#{h0xKKjijTdI},numeratorDescription,false,2,sqGRzCziswD
Indicator B,Uvn6LCg7dVU,CodeB,Indicator B description,Indicator B shortname,#{fbfJHSPpUQD},denominatorDescription,#{h0xKKjijTdI},numeratorDescription,false,2,sqGRzCziswD

Collection membership

In addition to importing objects, you can also choose to only import the group-member relationship between an object and a group. Currently, the following group and object pairs are supported

• Organisation Unit Group - Organisation Unit

• Data Element Group - Data Element

• Indicator Group - Indicator

• Option Group Set - Option Group

The CSV format for these imports are the same

Other objects

An example of category options looks like this:

name,uid,code,shortname
"Male",,"MALE"
"Female",,"FEMALE"

Deleted objects

The deleted objects resource provides a log of metadata objects being deleted.

/api/deletedObjects

Whenever an object of type metadata is deleted, a log is being kept of the uid, code, the type and the time of when it was deleted. This API is available at /api/deletedObjects field filtering and object filtering works similarly to other metadata resources.

Get deleted objects of type data elements:

GET /api/deletedObjects.json?klass=DataElement

Get deleted object of type indicator which was deleted in 2015 and forward:

GET /api/deletedObjects.json?klass=Indicator&deletedAt=2015-01-01

Favorites

Certain types of metadata objects can be marked as favorites for the currently logged in user. This applies currently for dashboards.

/api/dashboards/<uid>/favorite

To make a dashboard a favorite you can make a POST request (no content type required) to a URL like this:

/api/dashboards/iMnYyBfSxmM/favorite

To remove a dashboard as a favorite you can make a DELETE request using the same URL as above.

The favorite status will appear as a boolean favorite field on the object (e.g. the dashboard) in the metadata response.

Subscriptions

A logged user can subscribe to certain types of objects. Currently subscribable objects are those of type EventChart, EventReport, Map, Visualization and EventVisualization.

Note

The EventChart and EventReport objects are deprecated. Use EventVisualization instead.

To get the subscribers of an object (return an array of user IDs) you can make a GET request:

/api/<object-type>/<object-id>/subscribers

See example as follows:

/api/visualizations/DkPKc1EUmC2/subscribers

To check whether the current user is subscribed to an object (returns a boolean) you can perform a GET call:

/api/<object-type>/<object-id>/subscribed

See example as follows:

/api/visualizations/DkPKc1EUmC2/subscribed

To subscribe/de-subscribe to an object you perform a POST/DELETE request (no content type required):

/api/<object-type>/<object-id>/subscriber

File resources

File resources are objects used to represent and store binary content. The FileResource object itself contains the file meta-data (name, Content-Type, size, etc.) as well as a key allowing retrieval of the contents from a database-external file store. The FileResource object is stored in the database like any other but the content (file) is stored elsewhere and is retrievable using the contained reference (storageKey).

/api/fileResources

The contents of file resources are not directly accessible but are referenced from other objects (such as data values) to store binary content of virtually unlimited size.

To create a file resource that does not require a corresponding data value, POST to the endpoint /api/fileResources with a multipart upload:

curl "https://server/api/fileResources" -X POST
  -F "file=@/path/to/file/name-of-file.png"

curl "https://server/api/fileResources?uid=0123456789x" -X POST
  -F "file=@/path/to/file/name-of-file.png"

To create both a file resource and a data value that references the file, POST to the /api/dataValues/file endpoint in DHIS 2.36 or later:

curl "https://server/api/dataValues/file?de=xPTAT98T2Jd
  &pe=201301&ou=DiszpKrYNg8&co=Prlt0C1RF0s" -X POST
  -F "file=@/path/to/file/name-of-file.png"

For the api/fileResources endpoint, the only form parameter required is file, which is the file to upload. For the api/dataValues/file endpoint, the parameters required are the same as for a post to api/dataValues, with the addition of file.

The filename and content-type should also be included in the request but will be replaced with defaults when not supplied.

On successfully creating a file resource the returned data will contain a response field which in turn contains the fileResource like this:

{
  "httpStatus": "Accepted",
  "httpStatusCode": 202,
  "status": "OK",
  "response": {
    "responseType": "FileResource",
    "fileResource": {
      "name": "name-of-file.png",
      "created": "2015-10-16T16:34:20.654+0000",
      "lastUpdated": "2015-10-16T16:34:20.667+0000",
      "externalAccess": false,
      "publicAccess": "--------",
      "user": { ... },
      "displayName": "name-of-file.png",
      "contentType": "image/png",
      "contentLength": 512571,
      "contentMd5": "4e1fc1c3f999e5aa3228d531e4adde58",
      "storageStatus": "PENDING",
      "id": "xm4JwRwke0i"
    }
  }
}

Note that the response is a 202 Accepted, indicating that the returned resource has been submitted for background processing (persisting to the external file store in this case). Also, note the storageStatus field which indicates whether the contents have been stored or not. At this point, the persistence to the external store is not yet finished (it is likely being uploaded to a cloud-based store somewhere) as seen by the PENDING status.

Even though the content has not been fully stored yet the file resource can now be used, for example as referenced content in a data value (see Working with file data values). If we need to check the updated storageStatus or otherwise retrieve the metadata of the file, the fileResources endpoint can be queried.

curl "https://server/api/fileResources/xm4JwRwke0i" -H "Accept: application/json"

This request will return the FileResource object as seen in the response of the above example.

File resource constraints

• File resources must be referenced (assigned) from another object in order to be persisted in the long term. A file resource which is created but not referenced by another object such as a data value is considered to be in staging. Any file resources which are in this state and are older than two hours will be marked for deletion and will eventually be purged from the system.

• The ID returned by the initial creation of the file resource is not retrievable from any other location unless the file resource has been referenced (in which the ID will be stored as the reference), so losing it will require the POST request to be repeated and a new object to be created. The orphaned file resource will be cleaned up automatically.

• File resource objects are immutable, meaning modification is not allowed and requires creating a completely new resource instead.

File resource blocklist

Certain types of files are blocked from being uploaded for security reasons.

The following content types are blocked.

The following file extensions are blocked.

Metadata versioning

This section explains the metadata versioning APIs.

• /api/metadata/version: This endpoint will return the current metadata version of the system on which it is invoked.

Get metadata version examples

Example: Get the current metadata version of this system

Request:

/api/metadata/version

Response:

{
  "name": "Version_4",
  "created": "2016-06-30T06:01:28.684+0000",
  "lastUpdated": "2016-06-30T06:01:28.685+0000",
  "externalAccess": false,
  "displayName": "Version_4",
  "type": "BEST_EFFORT",
  "hashCode": "848bf6edbaf4faeb7d1a1169445357b0",
  "id": "Ayz2AEMB6ry"
}

Example: Get the details of version with name "Version_2"

Request:

/api/metadata/version?versionName=Version_2

Response:

{
  "name": "Version_2",
  "created": "2016-06-30T05:59:33.238+0000",
  "lastUpdated": "2016-06-30T05:59:33.239+0000",
  "externalAccess": false,
  "displayName": "Version_2",
  "type": "BEST_EFFORT",
  "hashCode": "8050fb1a604e29d5566675c86d02d10b",
  "id": "SaNyhusVxBG"
}

• /api/metadata/version/history: This endpoint will return the list of all metadata versions of the system on which it is invoked.

Get the list of all metadata versions

Example: Get the list of all versions in this system

Request:

/api/metadata/version/history

Response:

{
  "metadataversions": [{
    "name": "Version_1",
    "type": "BEST_EFFORT",
    "created": "2016-06-30T05:54:41.139+0000",
    "id": "SjnhUp6r4hG",
    "hashCode": "fd1398ff7ec9fcfd5b59d523c8680798"
  }, {
    "name": "Version_2",
    "type": "BEST_EFFORT",
    "created": "2016-06-30T05:59:33.238+0000",
    "id": "SaNyhusVxBG",
    "hashCode": "8050fb1a604e29d5566675c86d02d10b"
  }, {
    "name": "Version_3",
    "type": "BEST_EFFORT",
    "created": "2016-06-30T06:01:23.680+0000",
    "id": "FVkGzSjAAYg",
    "hashCode": "70b779ea448b0da23d8ae0bd59af6333"
  }]
}

Example: Get the list of all versions in this system created after "Version_2"

Request:

/api/metadata/version/history?baseline=Version_2

Response:

{
  "metadataversions": [{
    "name": "Version_3",
    "type": "BEST_EFFORT",
    "created": "2016-06-30T06:01:23.680+0000",
    "id": "FVkGzSjAAYg",
    "hashCode": "70b779ea448b0da23d8ae0bd59af6333"
  }, {
    "name": "Version_4",
    "type": "BEST_EFFORT",
    "created": "2016-06-30T06:01:28.684+0000",
    "id": "Ayz2AEMB6ry",
    "hashCode": "848bf6edbaf4faeb7d1a1169445357b0"
  }]
 }

• /api/metadata/version/create: This endpoint will create the metadata version for the version type as specified in the parameter.

Users can select the type of metadata which needs to be created. Metadata Version type governs how the importer should treat the given version. This type will be used while importing the metadata. There are two types of metadata.

• BEST_EFFORT: This type suggests that missing references can be ignored and the importer can continue importing the metadata (e.g. missing data elements on a data element group import).

• ATOMIC: This type ensures a strict type checking of the metadata references and the metadata import will fail if any of the references do not exist.

Note

It's recommended to have an ATOMIC type of versions to ensure that all systems (central and local) have the same metadata. Any missing reference is caught in the validation phase itself. Please see the importer details for a full explanation.

Create metadata version

Example: Create metadata version of type BEST_EFFORT

Request:

curl -X POST -u admin:district "https://play.dhis2.org/dev/api/metadata/version/create?type=BEST_EFFORT"

Response:

{
  "name": "Version_1",
  "created": "2016-06-30T05:54:41.139+0000",
  "lastUpdated": "2016-06-30T05:54:41.333+0000",
  "externalAccess": false,
  "publicAccess": "--------",
  "user": {
    "name": "John Traore",
    "created": "2013-04-18T17:15:08.407+0000",
    "lastUpdated": "2016-04-06T00:06:06.571+0000",
    "externalAccess": false,
    "displayName": "John Traore",
    "id": "xE7jOejl9FI"
  },
  "displayName": "Version_1",
  "type": "BEST_EFFORT",
  "hashCode": "fd1398ff7ec9fcfd5b59d523c8680798",
  "id": "SjnhUp6r4hG"
}

• /api/metadata/version/{versionName}/data: This endpoint will download the actual metadata specific to the version name passed as path parameter.

• /api/metadata/version/{versionName}/data.gz: This endpoint will download the actual metadata specific to the version name passed as path parameter in a compressed format (gzipped).

Download version metadata

Example: Get the actual metadata for "Version 5"

Request:

curl -u admin:district "https://play.dhis2.org/dev/api/metadata/version/Version_5/data"

Response:

{
  "date": "2016-06-30T06:10:23.120+0000",
  "dataElements": [
    {
      "code": "ANC 5th Visit",
      "created": "2016-06-30T06:10:09.870+0000",
      "lastUpdated": "2016-06-30T06:10:09.870+0000",
      "name": "ANC 5th Visit",
      "id": "sCuZKDsix7Y",
      "shortName": "ANC 5th Visit ",
      "aggregationType": "SUM",
      "domainType": "AGGREGATE",
      "zeroIsSignificant": false,
      "valueType": "NUMBER",
      "categoryCombo": {
        "id": "p0KPaWEg3cf"
      },
      "user": {
        "id": "xE7jOejl9FI"
      }
    }
  ]
}

Metadata synchronization

This section explains the Metadata Synchronization API available starting 2.24

• /api/metadata/sync: This endpoint performs metadata sync of the version name passed in the query parameter by downloading and importing the specified version from the remote server as defined in the settings app.

• This API should be used with utmost care. Please note that there is an alternate way to achieve sync in a completely automated manner by leveraging the Metadata Sync Task from the "Data Administration" app. See Chapter 22, Section 22.17 of User Manual for more details regarding Metadata Sync Task.

• This sync API can alternatively be used to sync metadata for the versions which have failed from the metadata sync scheduler. Due to its dependence on the given metadata version number, care should be taken for the order in which this gets invoked. E.g. If this api is used to sync some higher version from the central instance, then the sync might fail as the metadata dependencies are not present in the local instance.

• Assume the local instance is at Version_12 and if this endpoint is used to sync Version_15 (of type BEST_EFFORT) from the central instance, the scheduler will start syncing metadata from Version_16. So the local instance will not have the metadata versions between Version_12 and Version_15. You need to manually sync the missing versions using these endpoints only.

Sync metadata version

Example: Sync Version_6 from central system to this system

Request:

curl -u admin:district "https://play.dhis2.org/dev/api/metadata/sync?versionName=Version_6"

Metadata repository

DHIS2 provides a metadata repository containing metadata packages with various content. A metadata package is a DHIS2-compliant JSON document which describes a set of metadata objects.

To retrieve an index over available metadata packages you can issue a GET request to the metadataRepo resource:

GET /api/synchronization/metadataRepo

A metadata package entry contains information about the package and a URL to the relevant package. An index could look like this:

{
  "packages": [
    {
      "id": "sierre-leone-demo",
      "name": "Sierra Leone demo",
      "description": "Sierra Leone demo database",
      "version": "0.1",
      "href": "https://dhis2.org/metadata-repo/221/sierra-leone-demo/metadata.json"
    },
    {
      "id": "trainingland-org-units",
      "name": "Trainingland organisation units",
      "description": "Trainingland organisation units with four levels",
      "version": "0.1",
      "href": "https://dhis2.org/metadata-repo/221/trainingland-org-units/metadata.json"
    }
  ]
}

A client can follow the URLs and install a metadata package through a POST request with content type text/plain with the metadata package URL as the payload to the metadataPull resource:

POST /api/synchronization/metadataPull

An example curl command looks like this:

curl "localhost:8080/api/synchronization/metadataPull" -X POST
  -d "https://dhis2.org/metadata-repo/221/trainingland-org-units/metadata.json"
  -H "Content-Type:text/plain" -u admin:district

Reference to created by user

Each object created in DHIS2 will have a property named user which is linked to User who created the object.

From version 2.36 we have changed the name of this property to createdBy to avoid confusion.

However, in order to keep the backwards compability, the legacy user property is still included in the payload and works normally as before.

{
  "createdBy": {
      "displayName": "John Kamara",
      "name": "John Kamara",
      "id": "N3PZBUlN8vq",
      "username": "district"
  },
  "user": {
      "displayName": "John Kamara",
      "name": "John Kamara",
      "id": "N3PZBUlN8vq",
      "username": "district"
  }
}

Metadata proposal workflow

The metadata proposal workflow endpoint allows for a workflow of proposing and accepting changes to metadata.

/api/metadata/proposals

Propose a metadata change

A proposal always targets a single metadata object using:

POST /api/metadata/proposals

Depending on the payload the proposal could:

• Add a new metadata object.
• Update an existing metadata object references by ID.
• Remove an existing metadata object referenced by ID.

To propose adding a new metadata object send a JSON payload like the following:

{
  "type": "ADD",
  "target": "ORGANISATION_UNIT",
  "change": {"name":"My Unit", "shortName":"MyOU", "openingDate": "2020-01-01"}
}

To propose updating an existing metadata object send a JSON payload like in the below example:

{
  "type": "UPDATE",
  "target": "ORGANISATION_UNIT",
  "targetId": "<id>",
  "change": [
    {"op": "replace", "path": "/name", "value": "New name"}
  ]
}

To propose the removal of an existing object send a payload like in the last example:

{
  "type": "REMOVE",
  "target": "ORGANISATION_UNIT",
  "targetId": "<id>"
}

Only target type ORGANISATION_UNIT is supported currently.

Accept a metadata change proposal

To accept an open proposal use POST on the proposal resource

POST /api/metadata/proposals/<uid>

When successful the status of the proposal changes to status ACCEPTED. Once accepted the proposal can no longer be rejected.

Should a proposal fail to apply it changes to status NEEDS_UPDATE. The reason field contains a summary of the failures when this information is available.

Oppose a metadata change proposal

If a proposal isn't quite right and needs adjustment this can be indicated by opposing the proposal by sending a PATCH for the proposal resource

PATCH /api/metadata/proposals/<uid>

Optionally a plain text body can be added to this to give a reason why the proposal got opposed.

A opposed proposal must be in state PROPOSED and will change to state NEEDS_UPDATE.

Adjust a metadata change proposal

A proposal in state NEEDS_UPDATE needs to be adjusted before it can be accepted. To adjust the proposal a PUT request is made for the proposal's resource

PUT /api/metadata/proposals/<uid>

Such an adjustment can either be made without a body or with a JSON body containing an object with the updated change and targetId for the adjustment:

{
  "targetId": "<id>",
  "change": ...
}

Reject a metadata change proposal

To reject an open proposal use DELETE on the proposal resource

DELETE /api/metadata/proposals/<uid>

This changes the status of the proposal conclusively to REJECTED. No further changes can be made to this proposal. It is kept as a documentation of the events.

List metadata change proposals

All proposals can be listed:

GET /api/metadata/proposals/

The result list can be filtered using the filter parameter. For example, to list only accepted proposals use:

GET /api/metadata/proposals?filter=status:eq:ACCEPTED

Similarly to only show open proposals use:

GET /api/metadata/proposals?filter=status:eq:PROPOSED

Filters can also be applied to any field except change. Supported filter operators are those described in the Gist Metadata API. This also includes property transformers described for Gist API.

List of available fields are:

Viewing metadata change proposals

Individual change proposals can be viewed using

GET /api/metadata/proposals/<uid>

The fields parameter can be used to narrow the fields included for the shown object. For example:

GET /api/metadata/proposals/<uid>?fields=id,type,status,change

Metadata Attribute Value Type and validations

Metadata Gist API

The Metadata Gist API is a RESTful read-only JSON API to fetch and browse metadata. Items in this API contain the gist of the same item in the Metadata API.

The API is specifically designed to avoid:

• Large response payloads because of the inclusion of partial nested object graphs.
• Resource intensive in memory processing of requests (e.g. in memory filtering or object graph traversal).
• n + 1 database queries as a result of object graph traversal while rendering the response.

Comparison with Metadata API

The standard Metadata API is a flexible and powerful API built to serve any and every use case. The downside of this is that not all features and combinations can scale while keeping good performance in the presence of huge numbers of items. In particular lists with items where each item itself has a property which is a large collection of complex objects have proven problematic as they quickly reference a large part of the entire object graph.

The /gist API was added to provide a metadata API where scaling well is our first priority. The downside of this is that there are more distinct limits to what features are technically reasonable, which means not all features of the standard Metadata API exist for the Gist API.

The Gist API uses a divide and conquer strategy to avoid responses with large partial object graphs. Instead of including nested objects or lists it provides a /gist endpoint URI where this object or list can be viewed in isolation.

The /gist API refers to nested data using URIs rather than including it. This means if a client is interested in this nested information more requests are required but each of them is kept reasonable small and will scale well in context of huge number of potential items.

Known Differences:

• items only includes fields of referenced identifiable objects if these do not have an endpoint on their own
• it never includes identifiable collections of objects directly
• items by default do not include all available fields, but a subset that depends on context and parameters
• lists cannot be used without pager (therefore there is no pager parameter)
• fields with collections are not paged using the pager-transformer but through a paged API endpoint for the particular collection property
• items in a list, a collection property size or boolean transformer result always considers object sharing (the set of considered items is always the set visible to the user)
• Gist offers member(<id>) and not-member(<id>) collection field transformers
• Gist offers canRead and canWrite access check filter instead of filtering on the access property
• Gist offers using attribute UIDs as field and filter property names to allow listing or filtering based on custom attribute values
• Gist offers filter grouping

Known Limitations:

• by default only persisted are included; a handful of special non-persistent fields (synthetic fields) can be added explicitly; other non-persistent fields might be possible to extract using from transformation
• filters can only be applied to persisted fields
• orders can only be applied to persisted fields
• token filters are not available
• order is always case-sensitive
• pluck transformer limited to text properties
• fields which hold collections of simple (non-identifiable) items cannot always be included depending on how they are stored

Where possible to use the /gist API should be considered the preferable way of fetching metadata information.

Endpoints

The /gist API has 3 kinds of endpoints:

• /api/<object-type>/gist: paged list of all known and visible objects of the type (implicit auto=S)
• /api/<object-type>/<object-id>/gist: view single object by ID (implicit auto=L)
• /api/<object-type>/<object-id>/<field-name>/gist: paged list of all known and visible items in the collection of owner object's field (implicit auto=M; in case this is a simple field just the field value)

These endpoints correspond to the endpoints of the standard metadata API without the /gist suffix and share the majority of parameters and their options with that API.

Browsing Data

Since /gist API avoids deeply nested data structures in the response the details of referenced complex objects or list of objects is instead provided in form of a URI to the gist endpoint that only returns the complex object or list of objects. These URIs are provided by the apiEndpoints field of an item which is automatically added to an item when such references exist. The item property itself might contain a transformation result on the object or collection such as its size, emptiness, non-emptiness, id(s) or plucked property such as its name.

To manually browse data it can be handy to use the absoluteUrls=true parameter. Linkage between parts of the gist can now be followed directly in browsers that render JSON responses.

Parameters

All endpoints of the /gist API accept the same set of parameters. Parameters and their options that do not make sense in the endpoint context are ignored.

Overview

Parameters in alphabetical order:

The absoluteUrls Parameter

By default, URIs in apiEndpoints, href and the pager prev and next members are relative, starting with /<object-type>/ path.

The URIs can be changed to absolute URLs using the absoluteUrls parameter.

For example, /api/users/rWLrZL8rP3K/gist?fields=id,href returns:

{
  "id": "rWLrZL8rP3K",
  "href": "/users/rWLrZL8rP3K/gist"
}

whereas /api/users/rWLrZL8rP3K/gist?fields=id,href&absoluteUrls=true returns:

{
  "id": "rWLrZL8rP3K",
  "href": "http://localhost:8080/api/users/rWLrZL8rP3K/gist?absoluteUrls=true"
}

As the example shows the absoluteUrls parameter is also forwarded or carried over to the included URLs so allowing to browse the responses by following the provided URLs.

The auto Parameter

Each endpoint implicitly sets a default for the extent of fields matched by the * / :all fields selector:

• /api/<object-type>/gist: implies auto=S
• /api/<object-type>/<object-id>/gist: implies auto=L
• /api/<object-type>/<object-id>/<field-name>/gist: implies auto=M

The auto parameter is used to manually override the default to make list items include more or less fields. This setting again acts as a default which can be further overridden on a per field basis using an explicit transformation.

Possible options for auto are ("t-shirt sizes"):

• XS: includes only IDs and textual properties
• S: excludes complex (object) properties, collection are only linked (not counted)
• M: complex included as reference URL, references and collections as count and reference URL
• L: like M but references and collections included as IDs (OBS! unbound in size)
• XL: like L but references and collections included as ID objects: { "id": <id> }

For example, /api/users/gist would list items with fields id, surname, firstName, phoneNumber, email, lastUpdated whereas /api/users/gist?auto=XS only lists id, surname, firstName, phoneNumber, email. Using /api/users/gist?auto=L would also include organisationUnits, dataViewOrganisationUnits, teiSearchOrganisationUnits and userGroups each with the list of IDs of the members in the lists/sets.

The fields Parameter

Specifies the list of fields to include for each list item.

Fields are included in the result JSON objects for an item in the provided order. A preset in the list of fields is expanded to the fields it contains at the position in the fields list it appears. Fields within the preset are ordered from simple to complex.

If no fields parameter is provided fields=* is assumed. Note that the fields of the * preset also depend on the auto parameter

To remove a field use either !<name> or -<name> in the list of fields. For example to remove the userGroups from a user, use: