[HIVE-26230] Option to URL encode special chars in hbase.column.mapping that are valid HBase column family chars

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.0.0-alpha-2
• Labels: pull-request-available
• Assignee: Ádám Szita Assign to me
• Reporter: Ádám Szita
• Created: 16/May/22 13:18
• Updated: 16/Nov/22 13:50
• Resolved: 01/Jun/22 20:42

Description

HIVE-26015 and HIVE-26139 aimed to fix a missing special character handling of values provided for hbase.column.mapping. Values here are used as an URL for Ranger based authentication and special characters need to be URL encoded for this feature.

This is currently done only for # char. We should handle all special characters that are valid HBase column family characters but count as special characters for URLs.

The URL encoding of HIVE-26015 should come back, as in HBase we can have almost any characters for column family (excluding : / ). To make this a backward-compatible change, the URL encoding will essentially be optional, so users won't have to make changes to their working environment. Should they encounter a special character in their HBase table definition though, they can turn this URL encoding feature on, which in turn comes with the requirement from their end to update their Ranger policies so they are in URL encoded format for these tables.

Issue Links

• relates to: HIVE-26139 URL Encoding from HIVE-26015 was a bit too aggressive
• relates to: HIVE-26015 CREATE HBase table fails when SERDEPROPERTIES contain special characters
• links to: GitHub Pull Request #3314