eZ Systems Website
  Editor documentation

  Developer documentation

  Back to the top

The documentation is moving!

We're moving our developer docs to Github and a new site.
The migration is still in progress, so please contact us in #documentation-contrib on eZ Community Slack if you have any questions or feedback.

This space will not be updated from now on except for critical fixes.

Skip to end of metadata
Go to start of metadata

We will start by going through the various ways to find and retrieve content from eZ Platform using the API. While this will be covered in further dedicated documentation, it is necessary to explain a few basic concepts of the Public API. In the following recipes, you will learn about the general principles of the API as they are introduced in individual recipes.

Displaying values from a Content item

In this recipe, we will see how to fetch a Content item from the repository, and obtain its Field's content. 

Let's first see the full code. You can see the Command line version at

Viewing content

Let's analyze this code block by block.

This is the initialization part. As explained above, everything in the Public API goes through the repository via dedicated services. We get the repository from the service container, using the method get() of our container, obtained via $this->getContainer(). Using our $repository variable, we fetch the two services we will need using getContentService() and getFieldTypeService().

Everything starting from line 5 is about getting our Content and iterating over its Fields. You can see that the whole logic is part of a try/catch block. Since the Public API uses Exceptions for error handling, this is strongly encouraged, as it will allow you to conditionally catch the various errors that may happen. We will cover the exceptions we expect in a later paragraph.

The first thing we do is use the Content Service to load a Content item using its ID, 66: $contentService->loadContent ( 66 ). As you can see on the API doc page, this method expects a Content ID, and returns a Content Value Object.

This block is the one that actually displays the value.

It iterates over the Content item's Fields using the Content Type's FieldDefinitions ($contentType->fieldDefinitions).

For each Field Definition, we start by displaying its identifier ($fieldDefinition->identifier). We then get the Field Type instance using the Field Type Service ($fieldTypeService->getFieldType( $fieldDefinition->fieldTypeIdentifier )). This method expects the requested Field Type's identifier, as a string (ezstring, ezxmltext, etc.), and returns an eZ\Publish\API\Repository\FieldType object.

The Field Value object is obtained using the getFieldValue() method of the Content Value Object which we obtained using ContentService::loadContent().

Using the Field Type object, we can convert the Field Value to a hash using the toHash() method, provided by every Field Type. This method returns a primitive type (string, hash) out of a Field instance.

With this example, you should get a first idea on how you interact with the API. Everything is done through services, each service being responsible for a specific part of the repository (Content, Field Type, etc.).

Loading Content in different languages

Since we didn't specify any language code, our Field object is returned in the given Content item's main language.

If you want to take SiteAccess languages into account, you have two options:

  • V1.10 By providing prioritized languages on load() this will be taken into account by the returned Content object when retrieving translated properties like fields, for example:
    $contentService->loadContent( 66, $configResolver->getParameter('languages'));
    • Note: As of v2.0 this is planned to be done for you when you don't specify languages.
  • Or you can take advantage of TranslationHelpers as described in Content Rendering.

Otherwise if you want to use an altogether different language, you can specify a language code in the getField() call:

Exceptions handling

As said earlier, the Public API uses Exceptions to handle errors. Each method of the API may throw different exceptions, depending on what it does. Which exceptions can be thrown is usually documented for each method. In our case, loadContent() may throw two types of exceptions: NotFoundException, if the requested ID isn't found, and UnauthorizedException if the currently logged in user isn't allowed to view the requested content.

It is a good practice to cover each exception you expect to happen. In this case, since our Command takes the Content ID as a parameter, this ID may either not exist, or the referenced Content item may not be visible to our user. Both cases are covered with explicit error messages.

Traversing a Location subtree

This recipe will show how to traverse a Location's subtree. The full code implements a command that takes a Location ID as an argument and recursively prints this location's subtree.

In this code, we introduce the LocationService. This service is used to interact with Locations. We use two methods from this service: loadLocation() , and loadLocationChildren() .

Loading a Location

As for the ContentService, loadLocation() returns a Value Object, here a Location . Errors are handled with exceptions: NotFoundException if the Location ID couldn't be found, and  UnauthorizedException if the current repository user isn't allowed to view this Location.

Iterating over a Location's children

LocationService::loadLocationChildren() returns a LocationList Value Objects that we can iterate over.

Note that unlike loadLocation(), we don't need to care for permissions here: the currently logged-in user's permissions will be respected when loading children, and Locations that can't be viewed won't be returned at all.

Full code

Should you need more advanced children fetching methods, the  SearchService  is what you are looking for.

Viewing Content Metadata

Content is a central piece in the Public API. You will often need to start from a Content item, and dig in from its metadata. Basic content metadata is made available through ContentInfo objects. This Value Object mostly provides primitive fields: contentTypeId, publishedDate or mainLocationId. But it is also used to request further Content-related Value Objects from various services.

The full example implements an ezpublish:cookbook:view_content_metadata command that prints out all the available metadata, given a Content ID.

We introduce here several new services: URLAliasService , UserService and SectionService . The concept should be familiar to you now.

Services initialization

Setting the Repository User

In a command line script, the repository runs as if executed by the anonymous user. In order to identify it as a different user, you need to use the UserService as follows (in this example 14 is the ID of the administrator user):

This may be crucial when writing maintenance or synchronization scripts.

This is of course not required in template functions or controller code, as the HTTP layer will take care of identifying the user, and automatically set it in the repository.


Since v1.6.0, as the setCurrentUser method is deprecated, you need to use the following code (here for the admin user, to be replaced with a different login as needed):

The ContentInfo Value Object

We will now load a ContentInfo object using the provided ID and use it to get our Content item's metadata


Getting Content Locations

We first use  LocationService ::loadLocations() to get the Locations for our ContentInfo. This method returns an array of Location Value Objects. In this example, we print out the Location's path string (/path/to/content). We also use URLAliasService::reverseLookup() to get the Location's main URLAlias.



We now want to list relations from and to our Content. Since relations are versioned, we need to feed the ContentService::loadRelations() with a VersionInfo object. We can get the current version's VersionInfo using  ContentService::loadVersionInfo() . If we had been looking for an archived version, we could have specified the version number as the second argument to this method.

Browsing a Content's relations

We can iterate over the Relation objects array we got from loadRelations(), and use these Value Objects to get data about our relations. It has two main properties: destinationContentInfo, and sourceContentInfo. They also hold the relation type (embed, common, etc.), and the optional Field this relations is made with.

ContentInfo properties

We can of course get our Content item's metadata by using the Value Object's properties.

Primitive object metadata

Owning user

We can use UserService::loadUser() with Content ownerId property of our ContentInfo to load the Content's owner as a User Value Object.

To get the current version's creator, and not the content's owner, you need to use the creatorId property from the current version's VersionInfo object.


The Section's ID can be found in the sectionId property of the ContentInfo object. To get the matching Section Value Object, you need to use the SectionService::loadSection() method.


To conclude we can also iterate over the Content's version, as VersionInfo Value Objects.

We use the ContentService::loadVersions() method and get an array of VersionInfo objects.

Version statuses


From v1.11 you can use VersionInfo->isDraft(), VersionInfo->isArchived(), and VersionInfo->isPublished() to check version status. This replaces using STATUS_ constants as the recommended way to do it. For example:

Previously: if ($content->getVersionInfo()->status === VersionInfo::STATUS_DRAFT)
Now: if ($content->getVersionInfo()->isDraft())

The new methods also enable you to use the following in Twig templates:

  • version.draft

  • version.published

  • version.archived

In this section we will cover how the  SearchService  can be used to search for Content, by using a Query  and a combinations of Criteria you will get a  SearchResult  object back containing list of Content and count of total hits. In the future this object will also include facets, spell checking and "more like this" when running on a backend that supports it (for instance Solr).

Difference between filter and query

Query object contains two properties you can set criteria on, filter and query, and while you can mix and match use and use both at the same time, there is one distinction between the two:

  • query: Has an effect on scoring (relevancy) calculation, and thus also on the default sorting if no sortClause is specified, when used with Solr and Elastic.
    • Typically you'll use this for FullText search criterion, and otherwise place everything else on filter.

Performing a simple full text search

In this recipe, we will run a simple full text search over every compatible attribute.

Query and Criterion objects

We introduce here a new object: Query. It is used to build up a Content query based on a set of Criterion objects.



Multiple criteria can be grouped together using "logical criteria", such as LogicalAnd or LogicalOr. Since in this case we only want to run a text search, we simply use a FullText criterion object.

The full list of criteria can be found on your installation in the following directory vendor/ezsystems/ezpublish-kernel/eZ/Publish/API/Repository/Values/Content/Query/Criterion. Additionally you may look at integration tests like vendor/ezsystems/ezpublish-kernel/eZ/Publish/API/Repository/Tests/SearchServiceTest.php for more details on how these are used.


Running the search query and using the results

The Query object is given as an argument to SearchService::findContent() . This method returns a SearchResult object. This object provides you with various information about the search operation (number of results, time taken, spelling suggestions, or facets, as well as, of course, the results themselves.

The searchHits properties of the SearchResult object is an array of SearchHit objects. In valueObject property of SearchHit, you will find the Content object that matches the given Query.


 If you you are searching using a unique identifier, for instance using the Content ID or Content remote ID criterion, then you can use  SearchService::findSingle() , this takes a Criterion and returns a single Content item, or throws a NotFound exception if none is found.


Retrieving Sort Clauses for parent location


You can use the method $parentLocation->getSortClauses() to return an array of Sort Clauses for direct use on LocationQuery->sortClauses.


Performing an advanced search

As explained in the previous chapter, Criterion objects are grouped together using logical criteria. We will now see how multiple criteria objects can be combined into a fine grained search Query.

A Subtree criterion limits the search to the subtree with pathString, which looks like: /1/2/. A ContentTypeId Criterion to limit the search to Content of Content Type 1. Those two criteria are grouped with a LogicalAnd operator. The query is executed as before, with SearchService::findContent().

Fine-tuning search results


The $languageFilter parameter provides a prioritized list of languages for the current SiteAccess. Passing it is recommended for front-end use, because otherwise all languages of the Content items will be returned.

Additionally, you can make use of the useAlwaysAvailable argument of the $languageFilter. This in turn uses the alwaysAvailable flag whose default is set on Content Type. When it is set to true, it ensures that when a language from the prioritized list can't be matched, the Content will be returned in its main language.


Criterion\Visibility enables you to ensure that only visible content will be returned.

Note that the criterion behaves differently depending on the method you use, because Locations have visibility, but Content does not. This means that when using the LocationQuery (findLocations($query)), the method will return the Location, if it is visible. When used with the Query (findContent($query)), however, the Content item will be returned even if one of its Locations is visible (although others may be hidden). That is why using Criterion\Visibility is recommended with LocationQuery.

This example shows the usage of both $languageFilter and Criterion\Visibility:

Performing a fetch like search

A search isn't only meant for searching, it also provides the interface for what was called "fetch" in eZ Publish 4.x. As this is totally back-end agnostic, eZ Publish's "ezfind" fetch functions are now powered by Solr (or ElasticSearch in experimental, unsupported setups).

Following the examples above we now change it a bit to combine several criteria with both an AND and an OR condition.

ParentLocationId  criterion limits the search to the children of location 2. An array of "ContentTypeId" Criteria to limit the search to Content of ContentType's with id 1 or 2 grouped in a LogicalOr operator. Those two criteria are grouped with a LogicalAnd operator. As always the query is executed as before, with SearchService::findContent().

Want to do a subtree filter? Change the location filter to use the Subtree criterion filter as shown in the advanced search example above.


Using in() instead of OR

The above example is fine, but it can be optimized a bit by taking advantage of the fact that all filter criteria support being given an array of values (IN operator) instead of a single value (EQ operator).

You can also use the ContentTypeIdentifier  Criterion:


All filter criteria are capable of doing an "IN" selection, the ParentLocationId above could, for example, have been provided "array( 2, 43 )" to include second level children in both your content tree (2) and your media tree (43).

Performing a Faceted Search

Under construction

Faceted Search is not fully implemented yet,

  • Implemented Facets SOLR BUNDLE >=1.4: User, ContentType, & Section , see:   EZP-26465 - Search Facets M1 Closed
  • Not Implemented Facets: CriterionFacet, DateRangeFacet, FieldFacet, FieldRangeFacet, LocationFacet (meant for Location search), & TermFacet

You can register custom facet builder visitors with Solr for Content(Info) and SOLR BUNDLE >=1.4 Location search.

Contribution starting point

The link above is also the starting point for contributing visitors for other API FacetBuilders and Facets . As for integration tests, fixtures that will need adjustments are found in ezpublish-kernel , and those missing in that link but defined in SearchServiceTest, are basically not implemented yet.

To be able to take advantage of facets, we can set the Query->facetBuilders property, which will result in relevant facets being returned on SearchResult->facets. All facet builders share the following properties:

  • name: Recommended, to set the human readable name of the returned facet for use in UI, so if you need translation this value should already be translated.
  • minCount: Optional, the minimum of hits of a given grouping, e.g. minimum number of content items in a given facet for it to be returned
  • limit: Optional, Maximum number of facets to be returned; only X number of facets with the greatest number of hits will be returned.

As an example, let's apply UserFacet to be able to group content according to the creator:

Performing a pure search count

In many cases you might need the number of Content items matching a search, but with no need to do anything else with the results.

Thanks to the fact that the " searchHits " property of the SearchResult object always refers to the total amount, it is enough to run a standard search and set $limit to 0. This way no results will be retrieved, and the search will not be slowed down, even when the number of matching results is huge.




In this topic: