[SPARK-30001] can't lookup v1 tables whose names specify the session catalog

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 3.0.0
• Component/s: SQL
• Assignee: Terry Kim
• Reporter: Wenchen Fan
• Created: 22/Nov/19 12:14
• Updated: 17/Dec/19 04:44
• Resolved: 06/Dec/19 07:46

Description

A simple way to reproduce it

scala> sql("create table t using hive as select 1 as i")
res2: org.apache.spark.sql.DataFrame = []

scala> sql("select * from t").show
+---+
|  i|
+---+
|  1|
+---+

scala> sql("select * from spark_catalog.t").show
org.apache.spark.sql.AnalysisException: Table or view not found: spark_catalog.t; line 1 pos 14;
'Project [*]
+- 'UnresolvedRelation [spark_catalog, t]

The reason is that, we first go into `ResolveTables`, which lookups the table successfully, but then give up because it's a v1 table. Next we go into `ResolveRelations`, which do not recognize catalog name at all.

Similar to https://issues.apache.org/jira/browse/SPARK-29966 , we should make `ResolveRelations` responsible for lookup both v1 and v2 tables from the session catalog, and correctly recognize catalog name.

Issue Links

• links to: GitHub Pull Request #26684