Hi,
I am having an issue with correlated sub-queries such as the following
SELECT * FROM P WHERE EXISTS (SELECT '*' FROM B WHERE P.ID = B.ID)
Both Beeline and Java JDBC client falied with the following message
Error: Error while compiling statement: FAILED: SemanticException [Error
10250]: Line 1:51 Invalid SubQuery expression 'ID': For Exists/Not
Exists operator SubQuery must be Correlated. (state=42000,code=10250)
However the query succeeds if the cross-referenced column is referred to
using lowercase. Note that all other references are case-insensitive and
I can mix the case in whichever way I want.
SELECT * FROM P WHERE EXISTS (SELECT '*' FROM B WHERE P.id = B.ID)
The expected behavior for me will be that all identifiers are
case-insensitive.
Is it a bug or am I missing something here?
HDFS version: 2.7.2, rb165c4fe8a74265c792ce23f546c64604acf0e41
Hive Version 2.0.0, r7f9f1fcb8697fb33f0edc2c391930a3728d247d7 with Derby
metastore
Hive JDBC driver: hive-jdbc-2.0.0.2.3.5.0-81-standalone.jar
Best regards,
/jack
