viirya commented on code in PR #36809:
URL: https://github.com/apache/spark/pull/36809#discussion_r898623736
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala:
##########
@@ -4345,32 +4340,42 @@ object ApplyCharTypePadding extends Rule[LogicalPlan] {
}
/**
- * Removes all [[TempResolvedColumn]]s in the query plan. This is the last
resort, in case some
- * rules in the main resolution batch miss to remove [[TempResolvedColumn]]s.
We should run this
- * rule right after the main resolution batch.
+ * The rule `ResolveAggregationFunctions` in the main resolution batch creates
+ * [[TempResolvedColumn]] in filter conditions and sort expressions to hold
the temporarily resolved
+ * column with `agg.child`. When filter conditions or sort expressions are
resolved,
+ * `ResolveAggregationFunctions` will turn [[TempResolvedColumn]] to
[[AttributeReference]] if it's
+ * inside aggregate functions or group expressions, or turn it to
[[UnresolvedAttribute]] otherwise,
+ * hoping other rules can resolve it.
+ *
+ * This rule runs after the main resolution batch, and can still hit
[[TempResolvedColumn]] if
+ * filter conditions or sort expressions are not resolved. When this happens,
there is no point to
Review Comment:
It reads like a contradiction here. `ResolveAggregationFunctions` will turn
`TempResolvedColumn` to `UnresolvedAttribute` and hope other rules can resolve
it.
So why `RemoveTempResolvedColumn` can hit `TempResolvedColumn`? If it is not
resolved, isn't it turned to `UnresolvedAttribute`?
--
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]
