Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
4.0.0-alpha-1
Description
Hive stores all name related value as lower case, such as db_name, tbl_name, col_name etc. But the compare of FieldSchema does not ignore the case of name, which may cause incorrect result of compare.
Bug Description:
Some computing engines are case sensitive for column name. For example, Spark will add a table property to save the column fields when creating a table, and will replace column fields with this property when fetching table fields.
When calling ALTER TABLE ... ADD COLUMNS, the compare of fields between old table and new table will be not expected, and the ADD COLUMNS operation will be cascaded to PARTITIONS, which is unnecessary and time consuming if the table has many partitions.
Attachments
Issue Links
- links to