dongjoon-hyun commented on a change in pull request #27643: [SPARK-30886][SQL]
Warn two-parameter TRIM/LTRIM/RTRIM function usages
URL: https://github.com/apache/spark/pull/27643#discussion_r381842297
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
##########
@@ -1837,13 +1839,19 @@ class Analyzer(
}
AggregateExpression(agg, Complete, isDistinct, filter)
// This function is not an aggregate function, just return the
resolved one.
- case other =>
- if (isDistinct || filter.isDefined) {
- failAnalysis("DISTINCT or FILTER specified, " +
- s"but ${other.prettyName} is not an aggregate function")
- } else {
- other
+ case other if (isDistinct || filter.isDefined) =>
+ failAnalysis("DISTINCT or FILTER specified, " +
+ s"but ${other.prettyName} is not an aggregate function")
+ case e: String2TrimExpression if arguments.size == 2 =>
+ if (trimWarningEnabled.get) {
+ log.warn("Two-parameter TRIM/LTRIM/RTRIM function
signatures are deprecated." +
+ s" Use SQL syntax
`${e.prettyName.toUpperCase(Locale.ROOT)}(trimStr" +
+ s" FROM str)` instead of `${e.prettyName}`. Please refer
SPARK-28093.")
+ trimWarningEnabled.set(false)
Review comment:
Hi, @maropu . Since `CheckAnalysis` received already resolved expressions,
there is no way to distinguish `SQL syntax TRIM(trimStr FROM str)` from Parser
and `function invocation TRIM(trimStr, str)` from here.
Also, `failAnalysis("DISTINCT or FILTER specified` also happens here instead
of `CheckAnalysis`.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
