This is an automated email from the ASF dual-hosted git repository.
alamb pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/datafusion.git
The following commit(s) were added to refs/heads/main by this push:
new b084aa4934 fix: Implement AggregateUDFImpl::reverse_expr for StringAgg
(#17165)
b084aa4934 is described below
commit b084aa4934a54c15c90748402cf6a418d1e74bb7
Author: Nuno Faria <[email protected]>
AuthorDate: Mon Sep 8 12:06:40 2025 +0100
fix: Implement AggregateUDFImpl::reverse_expr for StringAgg (#17165)
* fix: Implement AggregateUDFImpl::reverse_expr for StringAgg
* Add a test with two invocations of aggregateion
---------
Co-authored-by: Andrew Lamb <[email protected]>
---
datafusion/functions-aggregate/src/string_agg.rs | 4 ++
datafusion/sqllogictest/test_files/aggregate.slt | 53 +++++++++++++++++++++++-
2 files changed, 56 insertions(+), 1 deletion(-)
diff --git a/datafusion/functions-aggregate/src/string_agg.rs
b/datafusion/functions-aggregate/src/string_agg.rs
index a3a040da3f..3986984b26 100644
--- a/datafusion/functions-aggregate/src/string_agg.rs
+++ b/datafusion/functions-aggregate/src/string_agg.rs
@@ -178,6 +178,10 @@ impl AggregateUDFImpl for StringAgg {
)))
}
+ fn reverse_expr(&self) -> datafusion_expr::ReversedUDAF {
+ datafusion_expr::ReversedUDAF::Reversed(string_agg_udaf())
+ }
+
fn documentation(&self) -> Option<&Documentation> {
self.doc()
}
diff --git a/datafusion/sqllogictest/test_files/aggregate.slt
b/datafusion/sqllogictest/test_files/aggregate.slt
index 35b2a6c03b..caf8d637ec 100644
--- a/datafusion/sqllogictest/test_files/aggregate.slt
+++ b/datafusion/sqllogictest/test_files/aggregate.slt
@@ -6203,6 +6203,58 @@ from t;
----
a,c,d,b
+# Test explain / reverse_expr for string_agg
+query TT
+explain select string_agg(k, ',' order by v) from t;
+----
+logical_plan
+01)Aggregate: groupBy=[[]], aggr=[[string_agg(t.k, Utf8(",")) ORDER BY [t.v
ASC NULLS LAST]]]
+02)--TableScan: t projection=[k, v]
+physical_plan
+01)AggregateExec: mode=Single, gby=[], aggr=[string_agg(t.k,Utf8(",")) ORDER
BY [t.v ASC NULLS LAST]]
+02)--SortExec: expr=[v@1 ASC NULLS LAST], preserve_partitioning=[false]
+03)----DataSourceExec: partitions=1, partition_sizes=[1]
+
+query T
+select string_agg(k, ',' order by v) from t;
+----
+c,a,b,d
+
+query TT
+explain select string_agg(k, ',' order by v desc) from t;
+----
+logical_plan
+01)Aggregate: groupBy=[[]], aggr=[[string_agg(t.k, Utf8(",")) ORDER BY [t.v
DESC NULLS FIRST]]]
+02)--TableScan: t projection=[k, v]
+physical_plan
+01)AggregateExec: mode=Single, gby=[], aggr=[string_agg(t.k,Utf8(",")) ORDER
BY [t.v DESC NULLS FIRST]]
+02)--SortExec: expr=[v@1 DESC], preserve_partitioning=[false]
+03)----DataSourceExec: partitions=1, partition_sizes=[1]
+
+query T
+select string_agg(k, ',' order by v desc) from t;
+----
+d,b,a,c
+
+# Call string_agg with both ASC and DESC orderings, and expect only one sort
+# (because the aggregate can handle reversed inputs)
+query TT
+explain select string_agg(k, ',' order by v asc), string_agg(k, ',' order by v
desc) from t;
+----
+logical_plan
+01)Aggregate: groupBy=[[]], aggr=[[string_agg(t.k, Utf8(",")) ORDER BY [t.v
ASC NULLS LAST], string_agg(t.k, Utf8(",")) ORDER BY [t.v DESC NULLS FIRST]]]
+02)--TableScan: t projection=[k, v]
+physical_plan
+01)AggregateExec: mode=Single, gby=[], aggr=[string_agg(t.k,Utf8(",")) ORDER
BY [t.v ASC NULLS LAST], string_agg(t.k,Utf8(",")) ORDER BY [t.v DESC NULLS
FIRST]]
+02)--SortExec: expr=[v@1 ASC NULLS LAST], preserve_partitioning=[false]
+03)----DataSourceExec: partitions=1, partition_sizes=[1]
+
+query TT
+select string_agg(k, ',' order by v asc), string_agg(k, ',' order by v desc)
from t;
+----
+c,a,b,d d,b,a,c
+
+
statement ok
drop table t;
@@ -7444,4 +7496,3 @@ NULL NULL
statement ok
drop table distinct_avg;
-
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
