Yibo Dong created SPARK-61970:
---------------------------------
Summary: Nested CHAR table field misses a short string literal
Key: SPARK-61970
URL: https://issues.apache.org/jira/browse/SPARK-61970
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 5.0.0
Environment: Spark 5.0.0-SNAPSHOT, upstream master at
\{{dada2a6db5d73b152d6ba16e4fe21406f86abfe3}} (2026-11-15), local mode, default
SQL settings
Reporter: Yibo Dong
h2. What happened
A nested CHAR field fails to match a short string literal, although the
equivalent top-level CHAR column matches it.
h2. How to reproduce
Run in a fresh Spark SQL session with a fresh warehouse. The last query is the
top-level control.
{code:sql}
CREATE TABLE nested_char (id INT, s STRUCT<c: CHAR(3)>) USING PARQUET;
INSERT INTO nested_char VALUES (1, named_struct('c', 'A')), (2,
named_struct('c', 'B'));
CREATE TABLE top_char (id INT, c CHAR(3)) USING PARQUET;
INSERT INTO top_char VALUES (1, 'A'), (2, 'B');
SELECT id FROM nested_char WHERE s.c = 'A' ORDER BY id;
SELECT id FROM top_char WHERE c = 'A' ORDER BY id;
{code}
h2. Expected result
The nested filter should return ID 1, just like the top-level control.
h2. Actual result
The nested filter returns no rows; the top-level control returns ID 1.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]