mikhailnik-db commented on code in PR #48748:
URL: https://github.com/apache/spark/pull/48748#discussion_r1840683697
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala:
##########
@@ -2216,21 +2216,25 @@ class Analyzer(override val catalogManager:
CatalogManager) extends RuleExecutor
numArgs: Int,
u: UnresolvedFunction): Expression = {
func match {
- case owg: SupportsOrderingWithinGroup if u.isDistinct =>
+ case owg: InverseDistributionFunction if u.isDistinct =>
throw
QueryCompilationErrors.distinctInverseDistributionFunctionUnsupportedError(
owg.prettyName)
- case owg: SupportsOrderingWithinGroup
+ case owg: InverseDistributionFunction
if !owg.orderingFilled && u.orderingWithinGroup.isEmpty =>
throw
QueryCompilationErrors.inverseDistributionFunctionMissingWithinGroupError(
owg.prettyName)
- case owg: SupportsOrderingWithinGroup
+ case owg: InverseDistributionFunction
Review Comment:
No, because I made `ListAgg` support any number of order expressions.
Plus this is a strange `case`, because the actual check of order expressions
number happens in `withOrderingWithinGroup(...)` implementation. This `case`
can match only `Mode` function if `mode(expr) within group (order by expr)`
syntax is used. And that's because `mode(expr)` is equivalent to `mode() within
group (order by expr)`, and here we actually just ban using two syntaxes
simultaneously. I think the code and error message is quite misleading.
`ListAgg` has a similar check below if `DISTINCT` is specified (see
`QueryCompilationErrors.functionAndOrderExpressionMismatchError`). It could be
reused for `mode`, but it requires behavior change and I don't think it's
worth it.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
