2020-08-10 21:58:57 +02:00

307 lines
8.9 KiB
Plaintext

[[elasticsearch.misc]]
= Miscellaneous Elasticsearch Operation Support
This chapter covers additional support for Elasticsearch operations that cannot be directly accessed via the repository interface. It is recommended to add those operations as custom implementation as described in <<repositories.custom-implementations>> .
[[elasticsearch.misc.filter]]
== Filter Builder
Filter Builder improves query speed.
====
[source,java]
----
private ElasticsearchOperations operations;
IndexCoordinates index = IndexCoordinates.of("sample-index");
SearchQuery searchQuery = new NativeSearchQueryBuilder()
.withQuery(matchAllQuery())
.withFilter(boolFilter().must(termFilter("id", documentId)))
.build();
Page<SampleEntity> sampleEntities = operations.searchForPage(searchQuery, SampleEntity.class, index);
----
====
[[elasticsearch.scroll]]
== Using Scroll For Big Result Set
Elasticsearch has a scroll API for getting big result set in chunks. This is internally used by Spring Data Elasticsearch to provide the implementations of the `<T> SearchHitsIterator<T> SearchOperations.searchForStream(Query query, Class<T> clazz, IndexCoordinates index)` method.
[source,java]
----
IndexCoordinates index = IndexCoordinates.of("sample-index");
SearchQuery searchQuery = new NativeSearchQueryBuilder()
.withQuery(matchAllQuery())
.withFields("message")
.withPageable(PageRequest.of(0, 10))
.build();
SearchHitsIterator<SampleEntity> stream = elasticsearchTemplate.searchForStream(searchQuery, SampleEntity.class, index);
List<SampleEntity> sampleEntities = new ArrayList<>();
while (stream.hasNext()) {
sampleEntities.add(stream.next());
}
stream.close();
----
There are no methods in the `SearchOperations` API to access the scroll id, if it should be necessary to access this, the following methods of the `ElasticsearchRestTemplate` can be used:
[source,java]
----
@Autowired ElasticsearchRestTemplate template;
IndexCoordinates index = IndexCoordinates.of("sample-index");
SearchQuery searchQuery = new NativeSearchQueryBuilder()
.withQuery(matchAllQuery())
.withFields("message")
.withPageable(PageRequest.of(0, 10))
.build();
SearchScrollHits<SampleEntity> scroll = template.searchScrollStart(1000, searchQuery, SampleEntity.class, index);
String scrollId = scroll.getScrollId();
List<SampleEntity> sampleEntities = new ArrayList<>();
while (scroll.hasSearchHits()) {
sampleEntities.addAll(scroll.getSearchHits());
scrollId = scroll.getScrollId();
scroll = template.searchScrollContinue(scrollId, 1000, SampleEntity.class);
}
template.searchScrollClear(scrollId);
----
To use the Scroll API with repository methods, the return type must defined as `Stream` in the Elasticsearch Repository. The implementation of the method will then use the scroll methods from the ElasticsearchTemplate.
[source,java]
----
interface SampleEntityRepository extends Repository<SampleEntity, String> {
Stream<SampleEntity> findBy();
}
----
[[elasticsearch.misc.sorts]]
== Sort options
In addition to the default sort options described <<repositories.paging-and-sorting>> Spring Data Elasticsearch has a `GeoDistanceOrder` class which can be used to have the result of a search operation ordered by geographical distance.
If the class to be retrieved has a `GeoPoint` property named _location_, the following `Sort` would sort the results by distance to the given point:
[source,java]
----
Sort.by(new GeoDistanceOrder("location", new GeoPoint(48.137154, 11.5761247)))
----
[[elasticsearch.misc.jointype]]
== Join-Type implementation
Spring Data Elasticsearch supports the https://www.elastic.co/guide/en/elasticsearch/reference/current/parent-join.html[Join data type] for creating the corresponding index mappings and for storing the relevant information.
=== Setting up the data
For an entity to be used in a parent child join relationship, it must have a property of type `JoinField` which must be annotated.
Let's assume a `Statement` entity where a statement may be a _question_, an _answer_, a _comment_ or a _vote_ (a _Builder_ is also shown in this example, it's not necessary, but later used in the sample code):
====
[source,java]
----
@Document(indexName = "statements")
public class Statement {
@Id
private String id;
@Field(type = FieldType.Text)
private String text;
@JoinTypeRelations(
relations =
{
@JoinTypeRelation(parent = "question", children = {"answer", "comment"}), <1>
@JoinTypeRelation(parent = "answer", children = "vote") <2>
}
)
private JoinField<String> relation; <3>
private Statement() {
}
public static StatementBuilder builder() {
return new StatementBuilder();
}
public String getId() {
return id;
}
public void setId(String id) {
this.id = id;
}
public String getText() {
return text;
}
public void setText(String text) {
this.text = text;
}
public JoinField<String> getRelation() {
return relation;
}
public void setRelation(JoinField<String> relation) {
this.relation = relation;
}
public static final class StatementBuilder {
private String id;
private String text;
private JoinField<String> relation;
private StatementBuilder() {
}
public StatementBuilder withId(String id) {
this.id = id;
return this;
}
public StatementBuilder withText(String text) {
this.text = text;
return this;
}
public StatementBuilder withRelation(JoinField<String> relation) {
this.relation = relation;
return this;
}
public Statement build() {
Statement statement = new Statement();
statement.setId(id);
statement.setText(text);
statement.setRelation(relation);
return statement;
}
}
}
----
<1> a question can have answers and comments
<2> an answer can have votes
<3> the `JoinField` property is used to combine the name (_question_, _answer_, _comment_ or _vote_) of the relation with the parent id. The generic type must be the same as the `@Id` annotated property.
====
Spring Data Elasticsearch will build the following mapping for this class:
====
[source,json]
----
{
"statements": {
"mappings": {
"properties": {
"_class": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword",
"ignore_above": 256
}
}
},
"relation": {
"type": "join",
"eager_global_ordinals": true,
"relations": {
"question": [
"answer",
"comment"
],
"answer": "vote"
}
},
"text": {
"type": "text"
}
}
}
}
}
----
====
=== Storing data
Given a repository for this class the following code inserts a question, two answers, a comment and a vote:
====
[source,java]
----
void init() {
repository.deleteAll();
Statement savedWeather = repository.save(
Statement.builder()
.withText("How is the weather?")
.withRelation(new JoinField<>("question")) <1>
.build());
Statement sunnyAnswer = repository.save(
Statement.builder()
.withText("sunny")
.withRelation(new JoinField<>("answer", savedWeather.getId())) <2>
.build());
repository.save(
Statement.builder()
.withText("rainy")
.withRelation(new JoinField<>("answer", savedWeather.getId())) <3>
.build());
repository.save(
Statement.builder()
.withText("I don't like the rain")
.withRelation(new JoinField<>("comment", savedWeather.getId())) <4>
.build());
repository.save(
Statement.builder()
.withText("+1 for the sun")
.withRelation(new JoinField<>("vote", sunnyAnswer.getId())) <5>
.build());
}
----
<1> create a question statement
<2> the first answer to the question
<3> the second answer
<4> a comment to the question
<5> a vote for the first answer
====
=== Retrieving data
Currently native search queries must be used to query the data, so there is no support from standard repository methods. <<repositories.custom-implementations>> can be used instead.
The following code shows as an example how to retrieve all entries that have a _vote_ (which must be _answers_, because only answers can have a vote) using an `ElasticsearchOperations` instance:
====
[source,java]
----
SearchHits<Statement> hasVotes() {
NativeSearchQuery query = new NativeSearchQueryBuilder()
.withQuery(hasChildQuery("vote", matchAllQuery(), ScoreMode.None))
.build();
return operations.search(query, Statement.class);
}
----
====