[HBASE-18067] Support a default converter for data read shell commands

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 2.0.0
• Component/s: shell
• Hadoop Flags: Reviewed
• Assignee: Josh Elser
• Reporter: Josh Elser
• Created: 17/May/17 21:42
• Updated: 18/Oct/18 15:02
• Resolved: 22/May/17 02:32

Description

The get and scan shell commands have the ability to specify some complicated syntax on how to encode the bytes read from HBase on a per-column basis. By default, bytes falling outside of a limited range of ASCII are just printed as hex.

It seems like the intent of these converts was to support conversion of certain numeric columns as a readable string (e.g. 1234).

However, if non-ascii encoded bytes are stored in the table (e.g. UTF-8 encoded bytes), we may want to treat all data we read as UTF-8 instead (e.g. if row+column+value are in Chinese). It would be onerous to require users to enumerate every column they're reading to parse as UTF-8 instead of the limited ascii range. We can provide an option to encode all values retrieved by the command.

Issue Links

• causes: HBASE-21178 [BC break] : Get and Scan operation with a custom converter_class not working