- Created by Dominika Kurek, last modified on May 24, 2016
You are viewing an old version of this page. View the current version.
Compare with Current View Page History
« Previous Version 6 Next »
Introduction
eZ Platform exposes very powerful Search API, allowing both Full Text search and querying the content repository using several built-in Criteria and Sort Clauses. These are supported across several search engines, allowing you to plug in another search engine without changing your code.
Available Search Engines
Currently 3 search engines exists on their own eZ Platform Bundles:
- Legacy, a database-powered search engines for basic needs.
- Solr, an integration providing better overall performance, much better scalability and support for more advance search capabilities RECOMMENDED
- ElasticSearch, similar to Solr integration, however currently not under active development EXPERIMENTAL, NOT SUPPORTED
Usage
Search Criteria and Sort Clauses
Search Criteria and Sort Clauses are value
object classes used for building Search Query, to define filter criteria and ordering of the result set. eZ Platform provides a number of standard Criteria and Sort Clauses that you can use out of the box and that should cover the majority of use cases.
Search Engine Handling of Search Criteria and Sort Clauses
As Search Criteria and Sort Clauses are value
objects which are used to define the Query from API perspective, they are are common for all storage engines. Each storage engine needs to implement its own handler
for corresponding Criterion and Sort Clause value
object, which will be used to translate the value object into storage specific search query.
Custom Criteria and Sort Clauses
Sometimes you will find that standard Search Criteria and Sort Clauses provided with eZ Publish are not sufficient for you needs. Most often this will be the case if you have developed a custom Field Type using external storage, which therefore can not be searched using standard Field Criterion.
On use of Field Criterion/SortClause with large databases
Field Criterion/SortClause does not perform well by design when using SQL database, so if you have a large database and want to use them you either need to use Solr search engine, or develop your own Custom Criterion / Sort Clause to avoid use of attributes (Fields) database table, and instead uses a custom simplified table which can handle the amount of data you have.
In this case you can implement a custom Criterion or Sort Clause, together with the corresponding handlers for the storage engine you are using.
Difference between Content and Location Search
These are two basic types of searches, you can either search for Locations or for Content. Each has dedicated methods in Search Service:
Type of search | Method in Search Service |
---|---|
Content | findContent() |
Content | findSingle() |
Location | findLocations() |
All Criteria and Sort Clauses will be accepted with Location Search, but not all of them can be used with Content Search. Reason for this is that while one Location always has exactly one Content item, one Content item can have multiple Locations. In this context some Criteria and Sort Clauses would produce ambiguous queries and such will therefore not be accepted by Content Search.
Content Search will explicitly refuse to accept Criteria and Sort Clauses implementing these abstract classes:
eZ\Publish\API\Repository\Values\Content\Query\Criterion\Location
eZ\Publish\API\Repository\Values\Content\SortClause\Criterion\Location
How to configure your own Criterion and Sort Clause Handlers
After you have implemented your Criterion / Sort Clause and its handler, you will need to configure the handler for the service container using dedicated service tags for each type of search. Doing so will automatically register it and handle your Criterion / Search Clause when it is given as a parameter to one of the Search Service methods.
Available tags for Criterion handlers in Legacy Storage Engine are:
ezpublish.search.legacy.gateway.criterion_handler.content
ezpublish.search.legacy.gateway.criterion_handler.location
Available tags for Sort Clause handlers in Legacy Storage Engine are:
ezpublish.search.legacy.gateway.sort_clause_handler.content
ezpublish.search.legacy.gateway.sort_clause_handler.location
You will find all the native handlers and the tags for the Legacy Storage Engine available in the eZ/Publish/Core/settings/storage_engines/legacy/search_query_handlers.yml file.
Example of registering ContentId Criterion handler, common for both Content and Location Search
Example of registering Depth Sort Clause handler for Location Search
See also
See also Symfony documentation about Service Container for passing parameters
Criteria Independence Example
- Location A is visible
- Location B is hidden
- Content has Location B
- Content is visible (it has Location A that is visible)
Reference
Search Criteria Reference
Criteria are the filters for Content and Location Search, for generic use of API Search see Search Criteria and Sort Clauses .
A Criterion consist of two parts just like SortClause and FacetBuilder:
- The API Value:
Criterion
- Specific handler per search engine:
Criterion Handler
Criterion
represents the value you use in the API, while CriterionHandler
deals with the business logic in the background translating the value to something the Search engine can understand.
Implementation and availability of a handler typically depends on search engine capabilities and limitations, currently only Legacy (SQL) Search Engine exists, and for instance its support for FullText and Field Criterion is not optimal and it is advised to avoid heavy use of these until future search engine arrives.
Common concepts for most Criteria
For how to use each and every Criterion see the list below, as it depends on the Criterion Value constructor, but in general you should be aware of the following common concepts:
target
: Exposed if the given Criterion supports targeting a specific sub field, example: FieldDefinition or Meta Data identifiervalue
: The value(s) to filter on, this is typically ascalar
orarray
ofscalars.
operator
: Exposed on some Criteria- All operators can be seen on
eZ\Publish\API\Repository\Values\Content\Query\Criterion\Operator
, at time of writing:IN, EQ, GT, GTE, LT, LTE, LIKE, BETWEEN, CONTAINS
- Most Criteria do not expose this and select
EQ
orIN
depending on if value isscalar
orarray
IN
&BETWEEN
always acts on anarray
of values, while the other operators act on singlescalar
value
- All operators can be seen on
-
valueData
: Additional value data, required by some Criteria, MapLocationDistance for instance
In the Legacy search engine, the field index/sort key column is limited to 255 characters by design.
Due to this storage limitation, searching content using the eZ Country Field Type when there are multiple countries selected may not return all the expected results.
List of Criteria
The list below reflects Criteria available in the eZ\Publish\API\Repository\Values\Content\Query\Criterion
namespace (it is also possible to make a custom Criterion):
Only for LocationSearch
Criterion | Constructor arguments description |
---|---|
Location\Depth | operator (IN, EQ, GT, GTE, LT, LTE, BETWEEN ), value being the Location depth(s) as an integer(s). |
Location\IsMainLocation | Whether or not the Location is a Main Locationvalue (Location\IsMainLocation::MAIN , Location\IsMainLocation::NOT_MAIN ). |
Location\Priority | Priorities are integers that can be used for sorting in ascending or descending order. What is higher or lower priority in relation to the priority number is left to the user choice.operator (GT, GTE, LT, LTE, BETWEEN ), value being the location priority(s) as an integer(s). |
Common
Criterion | Constructor arguments description |
---|---|
ContentId | value being scalar(s) representing the Content id. |
ContentTypeGroupId | value being scalar(s) representing the Content Typ eGroup id. |
ContentTypeId | value being scalar(s) representing the Content Type id. |
ContentTypeIdentifier | value being string(s) representing the Content Type identifier, example: "article". |
DateMetadata | target ( DateMetadata ::MODIFIED , DateMetadata ::CREATED),
being integer(s) representing unix timestamp. |
Field | |
FieldRelation | target ( FieldDefinition identifier), (
being array of scalars representing Content id of relation.Use of IN means relation needs to have one of the provided ID's, while CONTAINS implies it needs to have all provided id's. |
FullText |
|
LanguageCode | value being string(s) representing Language Code(s) on the Content (not on Fields), matchAlwaysAvailable as boolean. |
LocationId | value being scalar(s) representing the Location id. |
LocationRemoteId | value being string(s) representing the Location Remote id. |
LogicalAnd | A LogicalOperator that takes array of other Criteria, makes sure all Criteria match. |
LogicalNot | A LogicalOperator that takes array of other Criteria, makes sure none of the Criteria match. |
LogicalOr | A LogicalOperator that takes array of other Criteria, makes sure one of the Criteria match. |
MapLocationDistance | target ( FieldDefinition identifier), (
as float(s) from a position using latitude
as float,
as float as arguments |
MatchAll | No arguments, mainly for internal use when no filter or query is provided on Query object. |
MatchNone | No arguments, mainly for internal use by BlockingLimitation. |
ObjectStateId | value being string(s) representing the Content ObjectState id. |
ParentLocationId | value being scalar(s) representing the Parent's Location id |
RemoteId | value being string(s) representing the Content Remote id. |
SectionId | value being scalar(s) representing the Content Section id. |
Subtree | value being string(s) representing the Location id in which you can filter. If the Location Id is /1/2/20/42, you will filter everything under 42. |
UserMetadata | target (UserMetadata
::OWNER , UserMetadata
::GROUP ), being scalars (s) representing the User or User Group id(s). |
Visibility | value (Visibility
::VISIBLE , Visibility
::HIDDEN), Note: This acts on all assigned locations when used with ContentSearch, meaning hidden content will be returned if it has a location which is visible. Use LocationSearch to avoid this. |
Sort Clauses Reference
Sort Clauses are the sorting options for Content and Location Search in eZ Platform. For generic use of API Search see Search Criteria and Sort Clauses.
A Sort Clause consists of two parts just like Criterion and FacetBuilder:
- The API Value:
SortClause
- Specific handler per search engine:
SortClausesHandler
The SortClause
represents the value you use in the API, while SortClauseHandler
deals with the business logic in the background, translating the value to something the Search engine can understand.
Implementation and availability of a handler sometimes depends on search engine capabilities and limitations.
Common concepts for all Sort Clauses
For how to use each and every Sort Clause, see list below as it depends on the Sort Clause Value constructor, but in general you should be aware of the following common concept:
sortDirection
: The direction to perform the sort, eitherQuery::SORT_ASC
(default) orQuery:: SORT_DESC
List of Sort Clauses
The list below reflects Sort Clauses available in the eZ\Publish\API\Repository\Values\Content\Query\SortClause
namespace (it is also possible to make a custom Sort Clause):
Arguments starting with "?
" are optional.
Only for LocationSearch
Sort Clause | Constructor arguments description |
---|---|
Location\Depth | ?sortDirection |
Location\Id | ?sortDirection |
Location\IsMainLocation | ?sortDirection |
Location\Depth | ?sortDirection |
Location\Priority | ?sortDirection |
Location\Visibility | ?sortDirection |
Common
Sort Clause | Constructor arguments description |
---|---|
ContentId | ?sortDirection |
ContentName | ?sortDirection |
DateModified | ?sortDirection |
DatePublished
| ?sortDirection |
Field |
|
MapLocationDistance | typeIdentifier as string , fieldIdentifier as string ,
as float ,
as float, ? sortDirection as string |
SectionIdentifier | ?sortDirection |
SectionName | ?sortDirection |