Github user ajithme commented on the issue:
https://github.com/apache/spark/pull/22277
@jiangxb1987 Thanks you for the feedback. Couple of points
1. If introduce a predicate which refers to alias( as u mentioned a > z),
it will throw error
```
spark-sql> create table table1 (a int);
18/09/05 13:00:28 WARN HiveMetaStore: Location:
file:/user/hive/warehouse/table1 specified for non-external table:table1
Time taken: 0.152 seconds
spark-sql> select a, a as c from table1 where a > 10 and a > c;
18/09/05 13:01:04 WARN ObjectStore: Failed to get database global_temp,
returning NoSuchObjectException
Error in query: cannot resolve '`c`' given input columns: [table1.a]; line
1 pos 50;
'Project ['a, 'a AS c#6]
+- 'Filter ((a#7 > 10) && (a#7 > 'c))
+- SubqueryAlias table1
+- HiveTableRelation `default`.`table1`,
org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, [a#7]
```
So i think its invalid scenario for a > z.? please correct me if i am wrong
2) if we add a predicate like __a > a__ instead of __a > z__ ( self
referring) the PR still produces valid constrain list
```
(x#5 > x#5),(b#1 <=> y#6),(x#5 > 10),(z#7 <=> x#5),isnotnull(x#5)
```
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
