Marcelo Vanzin created SPARK-26330:
--------------------------------------
Summary: Duplicate query execution events generated for SQL
commands
Key: SPARK-26330
URL: https://issues.apache.org/jira/browse/SPARK-26330
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 2.4.0
Reporter: Marcelo Vanzin
Consider the following code:
{code:java}
spark.sql("create table foo (bar int)").show()
{code}
The command is executed eagerly (i.e. before {{show()}} is called) and
generates a query execution event. But when you call {{show()}}, a duplicate
event is generated, even though Spark does not execute anything at that point.
This can be a little more misleading when you do something like a CTAS, since
the duplicate events may cause listeners to think there were multiple inserts
when that's not true.
A fuller example that shows this (and you can look at the output that both
inputs to the listener are the same):
{code:java}
import org.apache.spark.sql.execution.QueryExecution
import org.apache.spark.sql.util.QueryExecutionListener
val lsnr = new QueryExecutionListener() {
override def onSuccess(funcName: String, qe: QueryExecution, durationNs:
Long): Unit = {
println(s"on success: $funcName -> ${qe.analyzed}")
}
override def onFailure(funcName: String, qe: QueryExecution, exception:
Exception): Unit = {
println(s"on failure: $funcName -> ${qe.analyzed}")
}
}
spark.sessionState.listenerManager.register(lsnr)
spark.sql("drop table if exists test")
val df = spark.sql("create table test(i int)")
df.show()
{code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
