26.5. Using the Infinispan Query DSL with Spark

The Infinispan Query DSL can be used as a filter for the InfinispanRDD, allowing data to be pre-filtered at the source rather than at RDD level.

Important

Data in the cache must have been encoded with protobuf for the querying DSL to function correctly. Instructions on using protobuf encoding are found in Section 17.5, “Protobuf Encoding”.

Consider the following example which retrieves a list of books that includes any author whose name contains Doe:

Example 26.9. Filtering by a Query (Scala)

import org.infinispan.client.hotrod.impl.query.RemoteQuery
import org.infinispan.client.hotrod.{RemoteCacheManager, Search}
import org.infinispan.spark.domain._
[...]
val query = Search.getQueryFactory(remoteCacheManager.getCache(getCacheName))
    .from(classOf[Book])
    .having("author").like("Doe")
    .toBuilder[RemoteQuery].build()
    
val rdd = createInfinispanRDD[Int, Book]
    .filterByQuery[Book]](query, classOf[Book])

import org.infinispan.client.hotrod.impl.query.RemoteQuery
import org.infinispan.client.hotrod.{RemoteCacheManager, Search}
import org.infinispan.spark.domain._
[...]
val query = Search.getQueryFactory(remoteCacheManager.getCache(getCacheName))
    .from(classOf[Book])
    .having("author").like("Doe")
    .toBuilder[RemoteQuery].build()
    
val rdd = createInfinispanRDD[Int, Book]
    .filterByQuery[Book]](query, classOf[Book])

Copy to Clipboard

Toggle word wrap

Projections are also fully supported; for instance, the above example may be adjusted to only obtain the title and publication year, and sorting on the latter field:

Example 26.10. Filtering with a Projection (Scala)

import org.infinispan.client.hotrod.impl.query.RemoteQuery
import org.infinispan.client.hotrod.{RemoteCacheManager, Search}
import org.infinispan.spark.domain._
[...]
val query = Search.getQueryFactory(remoteCacheManager.getCache(getCacheName))
    .select("title","publicationYear")
    .from(classOf[Book])
    .having("author").like("Doe")
    .groupBy("publicationYear")
    .toBuilder[RemoteQuery].build()
    
val rdd = createInfinispanRDD[Int, Book]
    .filterByQuery[Array[AnyRef]](query, classOf[Book])

import org.infinispan.client.hotrod.impl.query.RemoteQuery
import org.infinispan.client.hotrod.{RemoteCacheManager, Search}
import org.infinispan.spark.domain._
[...]
val query = Search.getQueryFactory(remoteCacheManager.getCache(getCacheName))
    .select("title","publicationYear")
    .from(classOf[Book])
    .having("author").like("Doe")
    .groupBy("publicationYear")
    .toBuilder[RemoteQuery].build()
    
val rdd = createInfinispanRDD[Int, Book]
    .filterByQuery[Array[AnyRef]](query, classOf[Book])

Copy to Clipboard

Toggle word wrap

In addition, if a filter has already been deployed to the JBoss Data Grid server it may be referenced by name, as seen below:

Example 26.11. Filtering with a Deployed Filter (Scala)

val rdd = InfinispanRDD[String,Book] = .... 
// "my-filter-factory" filter and converts Book to a String, and has two parameters
val filteredRDD = rdd.filterByCustom[String]("my-filter-factory", "param1", "param2")

val rdd = InfinispanRDD[String,Book] = .... 
// "my-filter-factory" filter and converts Book to a String, and has two parameters
val filteredRDD = rdd.filterByCustom[String]("my-filter-factory", "param1", "param2")

Copy to Clipboard

Toggle word wrap

Report a bug

返回顶部

此内容没有您所选择的语言版本。

26.5. Using the Infinispan Query DSL with Spark

学习

尝试、购买和销售

社区

关于红帽文档

让开源更具包容性

關於紅帽

Theme

Red Hat legal and privacy links

Red Hat legal and privacy links