Github user CodingCat commented on a diff in the pull request:
https://github.com/apache/spark/pull/19864#discussion_r155587468
--- Diff:
sql/core/src/test/scala/org/apache/spark/sql/execution/columnar/InMemoryColumnarQuerySuite.scala
---
@@ -479,4 +481,32 @@ class InMemoryColumnarQuerySuite extends QueryTest
with SharedSQLContext {
}
}
}
+
+ test("SPARK-22673: InMemoryRelation should utilize existing stats
whenever possible") {
+ withSQLConf("spark.sql.cbo.enabled" -> "true") {
+ // scalastyle:off
+ val workDir = s"${Utils.createTempDir()}/table1"
+ val data = Seq(100, 200, 300, 400).toDF("count")
+ data.write.parquet(workDir)
+ val dfFromFile = spark.read.parquet(workDir).cache()
+ val inMemoryRelation =
dfFromFile.queryExecution.optimizedPlan.collect {
+ case plan: InMemoryRelation => plan
+ }.head
+ // InMemoryRelation's stats is Long.MaxValue before the underlying
RDD is materialized
+ assert(inMemoryRelation.computeStats().sizeInBytes === Long.MaxValue)
+ // InMemoryRelation's stats is updated after materializing RDD
+ dfFromFile.collect()
+ assert(inMemoryRelation.computeStats().sizeInBytes === 16)
+ // test of catalog table
+ val dfFromTable = spark.catalog.createTable("table1",
workDir).cache()
+ val inMemoryRelation2 = dfFromTable.queryExecution.optimizedPlan.
+ collect { case plan: InMemoryRelation => plan }.head
+ // Even CBO enabled, InMemoryRelation's stats keeps as the default
one before table's stats
+ // is calculated
+ assert(inMemoryRelation2.computeStats().sizeInBytes ===
Long.MaxValue)
--- End diff --
done
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
