Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
3.5.0
Description
Using the temp view reproduces COUNT bug, returns nulls instead of 0.
With a table:
scala> spark.sql("""CREATE TABLE outer_table USING parquet AS SELECT * FROM VALUES
| (1, 1),
| (2, 1),
| (3, 3),
| (6, 6),
| (7, 7),
| (9, 9) AS inner_table(a, b)""")
val res6: org.apache.spark.sql.DataFrame = []
scala> spark.sql("CREATE TABLE null_table USING parquet AS SELECT CAST(null AS int) AS a, CAST(null as int) AS b ;")
val res7: org.apache.spark.sql.DataFrame = []
scala> spark.sql("""SELECT ( SELECT COUNT(null_table.a) AS aggAlias FROM null_table WHERE null_table.a = outer_table.a) FROM outer_table""").collect()
val res8: Array[org.apache.spark.sql.Row] = Array([0], [0], [0], [0], [0], [0])
With a view:
spark.sql("CREATE TEMPORARY VIEW outer_view(a, b) AS VALUES (1, 1), (2, 1),(3, 3), (6, 6), (7, 7), (9, 9);") spark.sql("CREATE TEMPORARY VIEW null_view(a, b) AS SELECT CAST(null AS int), CAST(null as int);") spark.sql("""SELECT ( SELECT COUNT(null_view.a) AS aggAlias FROM null_view WHERE null_view.a = outer_view.a) FROM outer_view""").collect() val res2: Array[org.apache.spark.sql.Row] = Array([null], [null], [null], [null], [null], [null])