linhongliu-db commented on a change in pull request #31286:
URL: https://github.com/apache/spark/pull/31286#discussion_r565768669
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
##########
@@ -1763,6 +1763,21 @@ class Analyzer(override val catalogManager:
CatalogManager)
def expandStarExpression(expr: Expression, child: LogicalPlan): Expression
= {
expr.transformUp {
case f1: UnresolvedFunction if containsStar(f1.arguments) =>
+ // SPECIAL CASE: We want to block count(table.*) because in spark,
count(table.*) will
+ // be expanded while count(*) will be converted to count(1). They
will produce different
+ // results and confuse users if there is any null values. For
count(t1.*, t2.*), it is
Review comment:
because expand `count(t1.*, t2.*)` is not ambiguous, I think no one will
argue that `count(t1.*, t2.*)` should equal `count(1)`. Also this usage is not
supported by the common database (MySQL, oracle, pgsql), so we are not in
conflict with others.
but `count(table.*)` is not same. spark SQL expands the columns, pgsql
converts it to `count(1)`, MySQL, Oracle (as well as ANSI) thinks this should
be disallowed.
So I think block `count(table.*)` can avoid ambiguous while keeping
`count(t1.*, t2.*)` can reduce the side effects.
----------------------------------------------------------------
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
