ibzib commented on a change in pull request #16200:
URL: https://github.com/apache/beam/pull/16200#discussion_r783400048
##########
File path:
sdks/java/extensions/sql/zetasql/src/main/java/org/apache/beam/sdk/extensions/sql/zetasql/translation/AggregateScanConverter.java
##########
@@ -240,18 +244,22 @@ private AggregateCall convertAggCall(
}
List<Integer> argList = new ArrayList<>();
- for (ResolvedExpr expr :
- ((ResolvedAggregateFunctionCall)
computedColumn.getExpr()).getArgumentList()) {
+ ResolvedAggregateFunctionCall expr = ((ResolvedAggregateFunctionCall)
computedColumn.getExpr());
+ List<ZetaSQLResolvedNodeKind.ResolvedNodeKind> resolvedNodeKinds =
+ Arrays.asList(RESOLVED_CAST, RESOLVED_COLUMN_REF,
RESOLVED_GET_STRUCT_FIELD);
+ for (int i = 0; i < expr.getArgumentList().size(); i++) {
// Throw an error if aggregate function's input isn't either a ColumnRef
or a cast(ColumnRef).
// TODO: is there a general way to handle aggregation calls conversion?
- if (expr.nodeKind() == RESOLVED_CAST
- || expr.nodeKind() == RESOLVED_COLUMN_REF
- || expr.nodeKind() == RESOLVED_GET_STRUCT_FIELD) {
+ ZetaSQLResolvedNodeKind.ResolvedNodeKind resolvedNodeKind =
+ expr.getArgumentList().get(i).nodeKind();
+ if (i == 0 && resolvedNodeKinds.contains(resolvedNodeKind)) {
Review comment:
Alright, for now let's do this:
- Throw an UnsupportedOperationException when literals are passed as the
first argument. In other words, keep the current behavior.
- Remove the new array_agg and timestamp test cases.
- We can add support for literals later if we think it's important. I filed
a separate JIRA: https://issues.apache.org/jira/browse/BEAM-13648
--
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]
