Hibernate.orgCommunity Documentation

Chapter 5. Querying

5.1. Building queries
5.1.1. Building a Lucene query using the Lucene API
5.1.2. Building a Lucene query with the Hibernate Search query DSL
5.1.3. Building a Hibernate Search query
5.2. Retrieving the results
5.2.1. Performance considerations
5.2.2. Result size
5.2.3. ResultTransformer
5.2.4. Understanding results
5.3. Filters
5.3.1. Using filters in a sharded environment
5.4. Optimizing the query process

The second most important capability of Hibernate Search is the ability to execute Lucene queries and retrieve entities managed by a Hibernate session. The search provides the power of Lucene without leaving the Hibernate paradigm, giving another dimension to the Hibernate classic search mechanisms (HQL, Criteria query, native SQL query).

Preparing and executing a query consists of four simple steps:

To access the querying facilities, you have to use a FullTextSession. This Search specific session wraps a regular org.hibernate.Session in order to provide query and indexing capabilities.

Example 5.1. Creating a FullTextSession

Session session = sessionFactory.openSession();

...

FullTextSession fullTextSession = Search.getFullTextSession(session);    

Once you have a FullTextSession you have two options to build the full-text query: the Hibernate Search query DSL or the native Lucene query.

If you use the Hibernate Search query DSL, it will look like this:

final QueryBuilder b = fullTextSession.getSearchFactory()
    .buildQueryBuilder().forEntity( Myth.class ).get();

org.apache.lucene.search.Query luceneQuery =
    b.keyword()
        .onField("history").boostedTo(3)
        .matching("storm")
        .createQuery();


org.hibernate.Query fullTextQuery = fullTextSession.createFullTextQuery( luceneQuery );

List result = fullTextQuery.list(); //return a list of managed objects    

You can alternatively write your Lucene query either using the Lucene query parser or Lucene programmatic API.

Example 5.2. Creating a Lucene query via the QueryParser

SearchFactory searchFactory = fullTextSession.getSearchFactory();
org.apache.lucene.queryParser.QueryParser parser = 
    new QueryParser("title", searchFactory.getAnalyzer(Myth.class) );
try {
    org.apache.lucene.search.Query luceneQuery = parser.parse( "history:storm^3" );
}
catch (ParseException e) {
    //handle parsing failure
}


org.hibernate.Query fullTextQuery = fullTextSession.createFullTextQuery(luceneQuery);

List result = fullTextQuery.list(); //return a list of managed objects    

Note

The Hibernate query built on top of the Lucene query is a regular org.hibernate.Query, which means you are in the same paradigm as the other Hibernate query facilities (HQL, Native or Criteria). The regular list() , uniqueResult(), iterate() and scroll() methods can be used.

In case you are using the Java Persistence APIs of Hibernate, the same extensions exist:

Example 5.3. Creating a Search query using the JPA API

EntityManager em = entityManagerFactory.createEntityManager();


FullTextEntityManager fullTextEntityManager = 

    org.hibernate.search.jpa.Search.getFullTextEntityManager(em);


...

final QueryBuilder b = fullTextEntityManager.getSearchFactory()

    .buildQueryBuilder().forEntity( Myth.class ).get();


org.apache.lucene.search.Query luceneQuery =

    b.keyword()

        .onField("history").boostedTo(3)

        .matching("storm")

        .createQuery();
javax.persistence.Query fullTextQuery = 
    fullTextEntityManager.createFullTextQuery( luceneQuery );


List result = fullTextQuery.getResultList(); //return a list of managed objects  

Note

The following examples we will use the Hibernate APIs but the same example can be easily rewritten with the Java Persistence API by just adjusting the way the FullTextQuery is retrieved.

5.1. Building queries

Hibernate Search queries are built on top of Lucene queries which gives you total freedom on the type of Lucene query you want to execute. However, once built, Hibernate Search wraps further query processing using org.hibernate.Query as your primary query manipulation API.

Using the Lucene API, you have several options. You can use the query parser (fine for simple queries) or the Lucene programmatic API (for more complex use cases). It is out of the scope of this documentation on how to exactly build a Lucene query. Please refer to the online Lucene documentation or get hold of a copy of Lucene In Action or Hibernate Search in Action.

Writing full-text queries with the Lucene programmatic API is quite complex. It's even more complex to understand the code once written. Besides the inherent API complexity, you have to remember to convert your parameters to their string equivalent as well as make sure to apply the correct analyzer to the right field (a ngram analyzer will for example use several ngrams as the tokens for a given word and should be searched as such).

The Hibernate Search query DSL makes use of a style of API called a fluent API. This API has a few key characteristics:

Let's see how to use the API. You first need to create a query builder that is attached to a given indexed entity type. This QueryBuilder will know what analyzer to use and what field bridge to apply. You can create several QueryBuilders (one for each entity type involved in the root of your query). You get the QueryBuilder from the SearchFactory.

QueryBuilder mythQB = searchFactory.buildQueryBuilder().forEntity( Myth.class ).get();

You can also override the analyzer used for a given field or fields. This is rarely needed and should be avoided unless you know what you are doing.

QueryBuilder mythQB = searchFactory.buildQueryBuilder()

    .forEntity( Myth.class )

        .overridesForField("history","stem_analyzer_definition")

    .get();

Using the query builder, you can then build queries. It is important to realize that the end result of a QueryBuilder is a Lucene query. For this reason you can easily mix and match queries generated via Lucene's query parser or Query objects you have assembled with the Lucene programmatic API and use them with the Hibernate Search DSL. Just in case the DSL is missing some features.

Let's start with the most basic use case - searching for a specific word:

Query luceneQuery = mythQB.keyword().onField("history").matching("storm").createQuery();

keyword() means that you are trying to find a specific word. onField() specifies in which Lucene field to look. matching() tells what to look for. And finally createQuery() creates the Lucene query object. A lot is going on with this line of code.

Let's see how you can search a property that is not of type string.

@Entity 

@Indexed 

public class Myth {

  @Field(index = Index.UN_TOKENIZED) 

  @DateBridge(resolution = Resolution.YEAR)

  public Date getCreationDate() { return creationDate; }

  public Date setCreationDate(Date creationDate) { this.creationDate = creationDate; }

  private Date creationDate;

  

  ...

}


Date birthdate = ...;

Query luceneQuery = mythQb.keywork().onField("creationDate").matching(birthdate).createQuery();

Note

In plain Lucene, you would have had to convert the Date object to its string representation (in this case the year).

This conversion works for any object, not just Date, provided that the FieldBridge has an objectToString method (and all built-in FieldBridge implementations do).

We make the example a little more advanced now and have a look at how to search a field that uses ngram analyzers. ngram analyzers index succession of ngrams of your words which helps to recover from user typos. For example the 3-grams of the word hibernate are hib, ibe, ber, rna, nat, ate.

@AnalyzerDef(name = "ngram",

  tokenizer = @TokenizerDef(factory = StandardTokenizerFactory.class ),

  filters = {

    @TokenFilterDef(factory = StandardFilterFactory.class),

    @TokenFilterDef(factory = LowerCaseFilterFactory.class),

    @TokenFilterDef(factory = StopFilterFactory.class),

    @TokenFilterDef(factory = NGramFilterFactory.class,

      params = { 

        @Parameter(name = "minGramSize", value = "3"),

        @Parameter(name = "maxGramSize", value = "3") } )

  }

)

@Entity 

@Indexed 

public class Myth {

  @Field(analyzer=@Analyzer(definition="ngram") 

  @DateBridge(resolution = Resolution.YEAR)

  public String getName() { return name; }

  public String setName(Date name) { this.name = name; }

  private String name;

  

  ...

}


Date birthdate = ...;

Query luceneQuery = mythQb.keyword().onField("name").matching("Sisiphus").createQuery();

The matching word "Sisiphus" will be lower-cased and then split into 3-grams: sis, isi, sip, phu, hus. Each of these n-gram will be part of the query. We will then be able to find the Sysiphus myth (with a y). All that is transparently done for you.

Note

If for some reason you do not want a specific field to use the field bridge or the analyzer you can call the ignoreAnalyzer() or ignoreFieldBridge() functions

To search for multiple possible words in the same field, simply add them all in the matching clause.

//search document with storm or lightning in their history

Query luceneQuery = 

    mythQB.keyword().onField("history").matching("storm lightning").createQuery();

To search the same word on multiple fields, use the onFields method.

Query luceneQuery = mythQB

    .keyword()

    .onFields("history","description","name")

    .matching("storm")

    .createQuery();

Sometimes, one field should be treated differently from another field even if searching the same term, you can use the andField() method for that.

Query luceneQuery = mythQB.keyword()

    .onField("history")

    .andField("name")

      .boostedTo(5)

    .andField("description")

    .matching("storm")

    .createQuery();

In the previous example, only field name is boosted to 5.

To execute a fuzzy query (based on the Levenshtein distance algorithm), start like a keyword query and add the fuzzy flag.

Query luceneQuery = mythQB

    .keyword()

      .fuzzy()

        .withThreshold( .8f )

        .withPrefixLength( 1 )

    .onField("history")

    .matching("starm")

    .createQuery();

threshold is the limit above which two terms are considering matching. It's a decimal between 0 and 1 and defaults to 0.5. prefixLength is the length of the prefix ignored by the "fuzzyness": while it defaults to 0, a non zero value is recommended for indexes containing a huge amount of distinct terms.

You can also execute wildcard queries (queries where some of parts of the word are unknown). ? represents a single character and * represents any character sequence. Note that for performance purposes, it is recommended that the query does not start with either ? or *.

Query luceneQuery = mythQB

    .keyword()

      .wildcard()

    .onField("history")

    .matching("sto*")

    .createQuery();

Note

Wildcard queries do not apply the analyzer on the matching terms. Otherwise the risk of * or ? being mangled is too high.

So far we have been looking for words or sets of words, you can also search exact or approximate sentences. Use phrase() to do so.

Query luceneQuery = mythQB

    .phrase()

    .onField("history")

    .matching("Thou shalt not kill")

    .createQuery();

You can search approximate sentences by adding a slop factor. The slop factor represents the number of other words permitted in the sentence: this works like a within or near operator

Query luceneQuery = mythQB

    .phrase()

      .withSlop(3)

    .onField("history")

    .matching("Thou kill")

    .createQuery();

After looking at all these query examples for searching for to a given word, it is time to introduce range queries (on numbers, dates, strings etc). A range query searches for a value in between given boundaries (included or not) or for a value below or above a given boundary (included or not).

//look for 0 <= starred < 3

Query luceneQuery = mythQB

    .range()

    .onField("starred")

    .from(0).to(3).excludeLimit()

    .createQuery();


//look for myths strictly BC

Date beforeChrist = ...;

Query luceneQuery = mythQB

    .range()

    .onField("creationDate")

    .below(beforeChrist).excludeLimit()

    .createQuery();

Finally, you can aggregate (combine) queries to create more complex queries. The following aggregation operators are available:

The subqueries can be any Lucene query including a boolean query itself. Let's look at a few examples:

//look for popular modern myths that are not urban
Date twentiethCentury = ...;
Query luceneQuery = mythQB
    .bool()
      .must( mythQB.keyword().onField("description").matching("urban").createQuery() )
        .not()
      .must( mythQB.range().onField("starred").above(4).createQuery() )
      .must( mythQB
        .range()
        .onField("creationDate")
        .above(twentiethCentury)
        .createQuery() )
    .createQuery();

//look for popular myths that are preferably urban
Query luceneQuery = mythQB
    .bool()
      .should( mythQB.keyword().onField("description").matching("urban").createQuery() )
      .must( mythQB.range().onField("starred").above(4).createQuery() )
    .createQuery();

//look for all myths except religious ones
Query luceneQuery = mythQB
    .all()
      .except( monthQb
        .keyword()
        .onField( "description_stem" 
        .matching( "religion" )
        .createQuery() 
      )
    .createQuery();

We already have seen several query options in the previous example, but lets summarize again the options for query types and fields:

Let's check out an example using some of these options

Query luceneQuery = mythQB

    .bool()

      .should( mythQB.keyword().onField("description").matching("urban").createQuery() )

      .should( mythQB

        .keyword()

        .onField("name")

          .boostedTo(3)

          .ignoreAnalyzer()

        .matching("urban").createQuery() )

      .must( mythQB

        .range()

          .boostedTo(5).withConstantScore()

        .onField("starred").above(4).createQuery() )

    .createQuery();

As you can see, the Hibernate Search query DSL is an easy to use and easy to read query API and by accepting and producing Lucene queries, you can easily incorporate query types not (yet) supported by the DSL. Please give us feedback!

So far we only covered the process of how to create your Lucene query (see Section 5.1, “Building queries”). However, this is only the first step in the chain of actions. Let's now see how to build the Hibernate Search query from the Lucene query.

Once the Lucene query is built, it needs to be wrapped into an Hibernate Query. If not specified otherwise, the query will be executed against all indexed entities, potentially returning all types of indexed classes.

Example 5.4. Wrapping a Lucene query into a Hibernate Query

FullTextSession fullTextSession = Search.getFullTextSession( session );

org.hibernate.Query fullTextQuery = fullTextSession.createFullTextQuery( luceneQuery );

It is advised, from a performance point of view, to restrict the returned types:

Example 5.5. Filtering the search result by entity type

fullTextQuery = fullTextSession.createFullTextQuery( luceneQuery, Customer.class );


// or


fullTextQuery = fullTextSession.createFullTextQuery( luceneQuery, Item.class, Actor.class );

In Example 5.5, “Filtering the search result by entity type” the first example returns only matching Customers, the second returns matching Actors and Items. The type restriction is fully polymorphic which means that if there are two indexed subclasses Salesman and Customer of the baseclass Person, it is possible to just specify Person.class in order to filter on result types.

Out of performance reasons it is recommended to restrict the number of returned objects per query. In fact is a very common use case anyway that the user navigates from one page to an other. The way to define pagination is exactly the way you would define pagination in a plain HQL or Criteria query.

Example 5.6. Defining pagination for a search query

org.hibernate.Query fullTextQuery = 

    fullTextSession.createFullTextQuery( luceneQuery, Customer.class );

fullTextQuery.setFirstResult(15); //start from the 15th element

fullTextQuery.setMaxResults(10); //return 10 elements

Tip

It is still possible to get the total number of matching elements regardless of the pagination via fulltextQuery.getResultSize()

Apache Lucene provides a very flexible and powerful way to sort results. While the default sorting (by relevance) is appropriate most of the time, it can be interesting to sort by one or several other properties. In order to do so set the Lucene Sort object to apply a Lucene sorting strategy.

Example 5.7. Specifying a Lucene Sort in order to sort the results



org.hibernate.search.FullTextQuery query = s.createFullTextQuery( query, Book.class );

org.apache.lucene.search.Sort sort = new Sort(new SortField("title"), SortField.STRING);
query.setSort(sort);

List results = query.list();

Tip

Be aware that fields used for sorting must not be tokenized (see Section 4.1.1.2, “@Field”).

When you restrict the return types to one class, Hibernate Search loads the objects using a single query. It also respects the static fetching strategy defined in your domain model.

It is often useful, however, to refine the fetching strategy for a specific use case.

Example 5.8. Specifying FetchMode on a query

Criteria criteria = 

    s.createCriteria( Book.class ).setFetchMode( "authors", FetchMode.JOIN );

s.createFullTextQuery( luceneQuery ).setCriteriaQuery( criteria );

In this example, the query will return all Books matching the luceneQuery. The authors collection will be loaded from the same query using an SQL outer join.

When defining a criteria query, it is not necessary to restrict the returned entity types when creating the Hibernate Search query from the full text session: the type is guessed from the criteria query itself.

Important

Only fetch mode can be adjusted, refrain from applying any other restriction.

Important

One cannot use setCriteriaQuery if more than one entity type is expected to be returned.

For some use cases, returning the domain object (including its associations) is overkill. Only a small subset of the properties is necessary. Hibernate Search allows you to return a subset of properties:

Example 5.9. Using projection instead of returning the full domain object

org.hibernate.search.FullTextQuery query = 
    s.createFullTextQuery( luceneQuery, Book.class );
query.setProjection( "id", "summary", "body", "mainAuthor.name" );
List results = query.list();
Object[] firstResult = (Object[]) results.get(0);
Integer id = firstResult[0];
String summary = firstResult[1];
String body = firstResult[2];
String authorName = firstResult[3];

Hibernate Search extracts the properties from the Lucene index and convert them back to their object representation, returning a list of Object[]. Projections avoid a potential database round trip (useful if the query response time is critical). However, it also has several constraints:

  • the properties projected must be stored in the index (@Field(store=Store.YES)), which increases the index size

  • the properties projected must use a FieldBridge implementing org.hibernate.search.bridge.TwoWayFieldBridge or org.hibernate.search.bridge.TwoWayStringBridge, the latter being the simpler version.

    Note

    All Hibernate Search built-in types are two-way.

  • you can only project simple properties of the indexed entity or its embedded associations. This means you cannot project a whole embedded entity.

  • projection does not work on collections or maps which are indexed via @IndexedEmbedded

Projection is also useful for another kind of use case. Lucene can provide metadata information about the results. By using some special projection constants, the projection mechanism can retrieve this metadata:

Example 5.10. Using projection in order to retrieve meta data

org.hibernate.search.FullTextQuery query = 
    s.createFullTextQuery( luceneQuery, Book.class );
query.setProjection( 
    FullTextQuery.SCORE, 
    FullTextQuery.THIS, 
    "mainAuthor.name" );
List results = query.list();
Object[] firstResult = (Object[]) results.get(0);
float score = firstResult[0];
Book book = firstResult[1];
String authorName = firstResult[2];

You can mix and match regular fields and projection constants. Here is the list of the available constants:

  • FullTextQuery.THIS: returns the initialized and managed entity (as a non projected query would have done).

  • FullTextQuery.DOCUMENT: returns the Lucene Document related to the object projected.

  • FullTextQuery.OBJECT_CLASS: returns the class of the indexed entity.

  • FullTextQuery.SCORE: returns the document score in the query. Scores are handy to compare one result against an other for a given query but are useless when comparing the result of different queries.

  • FullTextQuery.ID: the id property value of the projected object.

  • FullTextQuery.DOCUMENT_ID: the Lucene document id. Careful, Lucene document id can change overtime between two different IndexReader opening (this feature is experimental).

  • FullTextQuery.EXPLANATION: returns the Lucene Explanation object for the matching object/document in the given query. Do not use if you retrieve a lot of data. Running explanation typically is as costly as running the whole Lucene query per matching element. Make sure you use projection!

You can limit the time a query takes in Hibernate Search in two ways:

You can decide to stop a query if when it takes more than a predefined amount of time. Note that this is a best effort basis but if Hibernate Search still has significant work to do and if we are beyond the time limit, a QueryTimeoutException will be raised (org.hibernate.QueryTimeoutException or javax.persistence.QueryTimeoutException depending on your programmatic API).

To define the limit when using the native Hibernate APIs, use one of the following approaches

Example 5.11. Defining a timeout in query execution

Query luceneQuery = ...;
FullTextQuery query = fullTextSession.createFullTextQuery(luceneQuery, User.class);

//define the timeout in seconds
query.setTimeout(5);

//alternatively, define the timeout in any given time unit
query.setTimeout(450, TimeUnit.MILLISECONDS);

try {
    query.list();
}
catch (org.hibernate.QueryTimeoutException e) {
    //do something, too slow
}

Likewise getResultSize(), iterate() and scroll() honor the timeout but only until the end of the method call. That simply means that the methods of Iterable or the ScrollableResults ignore the timeout.

Note

explain() does not honor the timeout: this method is used for debug purposes and in particular to find out why a query is slow

When using JPA, simply use the standard way of limiting query execution time.

Example 5.12. Defining a timeout in query execution

Query luceneQuery = ...;
FullTextQuery query = fullTextEM.createFullTextQuery(luceneQuery, User.class);

//define the timeout in milliseconds
query.setHint( "javax.persistence.query.timeout", 450 );

try {
    query.getResultList();
}
catch (javax.persistence.QueryTimeoutException e) {
    //do something, too slow
}

Important

Remember, this is a best effort approach and does not guarantee to stop exactly on the specified timeout.

Alternatively, you can return the number of results which have already been fetched by the time the limit is reached. Note that only the Lucene part of the query is influenced by this limit. It is possible that, if you retrieve managed object, it takes longer to fetch these objects.

Warning

This approach is not compatible with the setTimeout approach.

To define this soft limit, use the following approach

Example 5.13. Defining a time limit in query execution

Query luceneQuery = ...;
FullTextQuery query = fullTextSession.createFullTextQuery(luceneQuery, User.class);

//define the timeout in seconds
query.limitExecutionTimeTo(500, TimeUnit.MILLISECONDS);
List results = query.list();

Likewise getResultSize(), iterate() and scroll() honor the time limit but only until the end of the method call. That simply means that the methods of Iterable or the ScrollableResults ignore the timeout.

You can determine if the results have been partially loaded by invoking the hasPartialResults method.

Example 5.14. Determines when a query returns partial results

Query luceneQuery = ...;
FullTextQuery query = fullTextSession.createFullTextQuery(luceneQuery, User.class);

//define the timeout in seconds
query.limitExecutionTimeTo(500, TimeUnit.MILLISECONDS);
List results = query.list();

if ( query.hasPartialResults() ) {
    displayWarningToUser();
}

If you use the JPA API, limitExecutionTimeTo and hasPartialResults are also available to you.

Warning

This approach is considered experimental

5.2. Retrieving the results

Once the Hibernate Search query is built, executing it is in no way different than executing a HQL or Criteria query. The same paradigm and object semantic applies. All the common operations are available: list(), uniqueResult(), iterate(), scroll().

If you expect a reasonable number of results (for example using pagination) and expect to work on all of them, list() or uniqueResult() are recommended. list() work best if the entity batch-size is set up properly. Note that Hibernate Search has to process all Lucene Hits elements (within the pagination) when using list() , uniqueResult() and iterate().

If you wish to minimize Lucene document loading, scroll() is more appropriate. Don't forget to close the ScrollableResults object when you're done, since it keeps Lucene resources. If you expect to use scroll, but wish to load objects in batch, you can use query.setFetchSize(). When an object is accessed, and if not already loaded, Hibernate Search will load the next fetchSize objects in one pass.

Important

Pagination is a preferred over scrolling.

It is sometime useful to know the total number of matching documents:

Of course it would be too costly to retrieve all the matching documents. Hibernate Search allows you to retrieve the total number of matching documents regardless of the pagination parameters. Even more interesting, you can retrieve the number of matching elements without triggering a single object load.

Example 5.15. Determining the result size of a query

org.hibernate.search.FullTextQuery query = 
    s.createFullTextQuery( luceneQuery, Book.class );
//return the number of matching books without loading a single one
assert 3245 == query.getResultSize(); 

org.hibernate.search.FullTextQuery query = 
    s.createFullTextQuery( luceneQuery, Book.class );
query.setMaxResult(10);
List results = query.list();
//return the total number of matching books regardless of pagination
assert 3245 == query.getResultSize();

Note

Like Google, the number of results is approximative if the index is not fully up-to-date with the database (asynchronous cluster for example).

As seen in Section 5.1.3.5, “Projection” projection results are returns as Object arrays. This data structure is not always matching the application needs. In this cases It is possible to apply a ResultTransformer which post query execution can build the needed data structure:

Example 5.16. Using ResultTransformer in conjunction with projections

org.hibernate.search.FullTextQuery query = 
    s.createFullTextQuery( luceneQuery, Book.class );
query.setProjection( "title", "mainAuthor.name" );

query.setResultTransformer( 
    new StaticAliasToBeanResultTransformer( 
        BookView.class, 
        "title", 
        "author" ) 
);
List<BookView> results = (List<BookView>) query.list();
for(BookView view : results) {
    log.info( "Book: " + view.getTitle() + ", " + view.getAuthor() );
}

Examples of ResultTransformer implementations can be found in the Hibernate Core codebase.

You will find yourself sometimes puzzled by a result showing up in a query or a result not showing up in a query. Luke is a great tool to understand those mysteries. However, Hibernate Search also gives you access to the Lucene Explanation object for a given result (in a given query). This class is considered fairly advanced to Lucene users but can provide a good understanding of the scoring of an object. You have two ways to access the Explanation object for a given result:

The first approach takes a document id as a parameter and return the Explanation object. The document id can be retrieved using projection and the FullTextQuery.DOCUMENT_ID constant.

Warning

The Document id has nothing to do with the entity id. Do not mess up these two notions.

The second approach let's you project the Explanation object using the FullTextQuery.EXPLANATION constant.

Example 5.17. Retrieving the Lucene Explanation object using projection

FullTextQuery ftQuery = s.createFullTextQuery( luceneQuery, Dvd.class )
        .setProjection( 
             FullTextQuery.DOCUMENT_ID, 
             FullTextQuery.EXPLANATION, 
             FullTextQuery.THIS );
@SuppressWarnings("unchecked") List<Object[]> results = ftQuery.list();
for (Object[] result : results) {
    Explanation e = (Explanation) result[1];
    display( e.toString() );
}

Be careful, building the explanation object is quite expensive, it is roughly as expensive as running the Lucene query again. Don't do it if you don't need the object

5.3. Filters

Apache Lucene has a powerful feature that allows to filter query results according to a custom filtering process. This is a very powerful way to apply additional data restrictions, especially since filters can be cached and reused. Some interesting use cases are:

Hibernate Search pushes the concept further by introducing the notion of parameterizable named filters which are transparently cached. For people familiar with the notion of Hibernate Core filters, the API is very similar:

Example 5.18. Enabling fulltext filters for a given query

fullTextQuery = s.createFullTextQuery( query, Driver.class );
fullTextQuery.enableFullTextFilter("bestDriver");
fullTextQuery.enableFullTextFilter("security").setParameter( "login", "andre" );
fullTextQuery.list(); //returns only best drivers where andre has credentials

In this example we enabled two filters on top of the query. You can enable (or disable) as many filters as you like.

Declaring filters is done through the @FullTextFilterDef annotation. This annotation can be on any @Indexed entity regardless of the query the filter is later applied to. This implies that filter definitions are global and their names must be unique. A SearchException is thrown in case two different @FullTextFilterDef annotations with the same name are defined. Each named filter has to specify its actual filter implementation.

Example 5.19. Defining and implementing a Filter

@Entity
@Indexed
@FullTextFilterDefs( {
    @FullTextFilterDef(name = "bestDriver", impl = BestDriversFilter.class), 
    @FullTextFilterDef(name = "security", impl = SecurityFilterFactory.class) 
})
public class Driver { ... }
public class BestDriversFilter extends org.apache.lucene.search.Filter {

    public DocIdSet getDocIdSet(IndexReader reader) throws IOException {
        OpenBitSet bitSet = new OpenBitSet( reader.maxDoc() );
        TermDocs termDocs = reader.termDocs( new Term( "score", "5" ) );
        while ( termDocs.next() ) {
            bitSet.set( termDocs.doc() );
        }
        return bitSet;
    }
}

BestDriversFilter is an example of a simple Lucene filter which reduces the result set to drivers whose score is 5. In this example the specified filter implements the org.apache.lucene.search.Filter directly and contains a no-arg constructor.

If your Filter creation requires additional steps or if the filter you want to use does not have a no-arg constructor, you can use the factory pattern:

Example 5.20. Creating a filter using the factory pattern

@Entity
@Indexed
@FullTextFilterDef(name = "bestDriver", impl = BestDriversFilterFactory.class)
public class Driver { ... }

public class BestDriversFilterFactory {

    @Factory
    public Filter getFilter() {
        //some additional steps to cache the filter results per IndexReader
        Filter bestDriversFilter = new BestDriversFilter();
        return new CachingWrapperFilter(bestDriversFilter);
    }
}

Hibernate Search will look for a @Factory annotated method and use it to build the filter instance. The factory must have a no-arg constructor.

Named filters come in handy where parameters have to be passed to the filter. For example a security filter might want to know which security level you want to apply:

Example 5.21. Passing parameters to a defined filter

fullTextQuery = s.createFullTextQuery( query, Driver.class );
fullTextQuery.enableFullTextFilter("security").setParameter( "level", 5 );

Each parameter name should have an associated setter on either the filter or filter factory of the targeted named filter definition.

Example 5.22. Using parameters in the actual filter implementation

public class SecurityFilterFactory {
    private Integer level;

    /**
     * injected parameter
     */
    public void setLevel(Integer level) {
        this.level = level;
    }

    @Key
    public FilterKey getKey() {
        StandardFilterKey key = new StandardFilterKey();
        key.addParameter( level );
        return key;
    }

    @Factory
    public Filter getFilter() {
        Query query = new TermQuery( new Term("level", level.toString() ) );
        return new CachingWrapperFilter( new QueryWrapperFilter(query) );
    }
}

Note the method annotated @Key returning a FilterKey object. The returned object has a special contract: the key object must implement equals() / hashCode() so that 2 keys are equal if and only if the given Filter types are the same and the set of parameters are the same. In other words, 2 filter keys are equal if and only if the filters from which the keys are generated can be interchanged. The key object is used as a key in the cache mechanism.

@Key methods are needed only if:

  • you enabled the filter caching system (enabled by default)

  • your filter has parameters

In most cases, using the StandardFilterKey implementation will be good enough. It delegates the equals() / hashCode() implementation to each of the parameters equals and hashcode methods.

As mentioned before the defined filters are per default cached and the cache uses a combination of hard and soft references to allow disposal of memory when needed. The hard reference cache keeps track of the most recently used filters and transforms the ones least used to SoftReferences when needed. Once the limit of the hard reference cache is reached additional filters are cached as SoftReferences. To adjust the size of the hard reference cache, use hibernate.search.filter.cache_strategy.size (defaults to 128). For advanced use of filter caching, you can implement your own FilterCachingStrategy. The classname is defined by hibernate.search.filter.cache_strategy.

This filter caching mechanism should not be confused with caching the actual filter results. In Lucene it is common practice to wrap filters using the IndexReader around a CachingWrapperFilter. The wrapper will cache the DocIdSet returned from the getDocIdSet(IndexReader reader) method to avoid expensive recomputation. It is important to mention that the computed DocIdSet is only cachable for the same IndexReader instance, because the reader effectively represents the state of the index at the moment it was opened. The document list cannot change within an opened IndexReader. A different/new IndexReader instance, however, works potentially on a different set of Documents (either from a different index or simply because the index has changed), hence the cached DocIdSet has to be recomputed.

Hibernate Search also helps with this aspect of caching. Per default the cache flag of @FullTextFilterDef is set to FilterCacheModeType.INSTANCE_AND_DOCIDSETRESULTS which will automatically cache the filter instance as well as wrap the specified filter around a Hibernate specific implementation of CachingWrapperFilter (org.hibernate.search.filter.CachingWrapperFilter). In contrast to Lucene's version of this class SoftReferences are used together with a hard reference count (see discussion about filter cache). The hard reference count can be adjusted using hibernate.search.filter.cache_docidresults.size (defaults to 5). The wrapping behaviour can be controlled using the @FullTextFilterDef.cache parameter. There are three different values for this parameter:

ValueDefinition
FilterCacheModeType.NONENo filter instance and no result is cached by Hibernate Search. For every filter call, a new filter instance is created. This setting might be useful for rapidly changing data sets or heavily memory constrained environments.
FilterCacheModeType.INSTANCE_ONLYThe filter instance is cached and reused across concurrent Filter.getDocIdSet() calls. DocIdSet results are not cached. This setting is useful when a filter uses its own specific caching mechanism or the filter results change dynamically due to application specific events making DocIdSet caching in both cases unnecessary.
FilterCacheModeType.INSTANCE_AND_DOCIDSETRESULTSBoth the filter instance and the DocIdSet results are cached. This is the default value.

Last but not least - why should filters be cached? There are two areas where filter caching shines:

  • the system does not update the targeted entity index often (in other words, the IndexReader is reused a lot)

  • the Filter's DocIdSet is expensive to compute (compared to the time spent to execute the query)

It is possible, in a sharded environment to execute queries on a subset of the available shards. This can be done in two steps:

Let's first look at an example of sharding strategy that query on a specific customer shard if the customer filter is activated.

public class CustomerShardingStrategy implements IndexShardingStrategy {

 // stored DirectoryProviders in a array indexed by customerID
 private DirectoryProvider<?>[] providers;
 
 public void initialize(Properties properties, DirectoryProvider<?>[] providers) {
  this.providers = providers;
 }

 public DirectoryProvider<?>[] getDirectoryProvidersForAllShards() {
  return providers;
 }

 public DirectoryProvider<?> getDirectoryProviderForAddition(Class<?> entity, Serializable id, String idInString, Document document) {
  Integer customerID = Integer.parseInt(document.getField("customerID").stringValue());
  return providers[customerID];
 }

 public DirectoryProvider<?>[] getDirectoryProvidersForDeletion(Class<?> entity, Serializable id, String idInString) {
  return getDirectoryProvidersForAllShards();
 }

 /**
  * Optimization; don't search ALL shards and union the results; in this case, we 
  * can be certain that all the data for a particular customer Filter is in a single
  * shard; simply return that shard by customerID.
  */
 public DirectoryProvider<?>[] getDirectoryProvidersForQuery(FullTextFilterImplementor[] filters) {
  FFullTextFilter filter = getCustomerFilter(filters, "customer");
  if (filter == null) {
   return getDirectoryProvidersForAllShards();
  }
  else {
   return new DirectoryProvider[] { providers[Integer.parseInt(filter.getParameter("customerID").toString())] };
  }
 }

 private FullTextFilter getFilter(FullTextFilterImplementor[] filters, String name) {
  for (FullTextFilterImplementor filter: filters) {
   if (filter.getName().equals(name)) return filter;
  }
  return null;
 }
}

In this example, if the filter named customer is present, we make sure to only use the shard dedicated to this customer. Otherwise, we return all shards. A given Sharding strategy can react to one or more filters and depends on their parameters.

The second step is simply to activate the filter at query time. While the filter can be a regular filter (as defined in Section 5.3, “Filters”) which also filters Lucene results after the query, you can make use of a special filter that will only be passed to the sharding strategy and otherwise ignored for the rest of the query. Simply use the ShardSensitiveOnlyFilter class when declaring your filter.

@Entity @Indexed
@FullTextFilterDef(name="customer", impl=ShardSensitiveOnlyFilter.class)
public class Customer {
   ...
}


FullTextQuery query = ftEm.createFullTextQuery(luceneQuery, Customer.class);
query.enableFulltextFilter("customer").setParameter("CustomerID", 5);
@SuppressWarnings("unchecked")
List<Customer> results = query.getResultList();

Note that by using the ShardSensitiveOnlyFilter, you do not have to implement any Lucene filter. Using filters and sharding strategy reacting to these filters is recommended to speed up queries in a sharded environment.

5.4. Optimizing the query process

Query performance depends on several criteria: