[SPARK-5278] check ambiguous reference to fields in Spark SQL is incompleted

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.3.0
• Component/s: SQL
• Assignee: Wenchen Fan
• Reporter: Wenchen Fan
• Created: 16/Jan/15 06:05
• Updated: 14/Aug/16 11:23
• Resolved: 06/Feb/15 21:11

Description

at hive context

for json string like

{"a": {"b": 1, "B": 2}}

The SQL `SELECT a.b from t` will report error for ambiguous reference to fields.
But for json string like

{"a": [{"b": 1, "B": 2}]}

The SQL `SELECT a[0].b from t` will pass and pick the first `b`

Issue Links

• is duplicated by: SPARK-6273 Got error when one table's alias name is the same with other table's column name
• links to: [Github] Pull Request #4068 (cloud-fan)