[HBASE-3623] Allow non-XML representable separator characters in the ImportTSV tool

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.90.1
• Fix Version/s: 0.92.0
• Component/s: mapreduce
• Labels: import

Environment

Cloudera Hadoop/HBase (3B4)

• Hadoop Flags: Reviewed
• Release Note: Allow use of non-XML friendly characters as separators in the ImportTSV tool.
• Tags: importtsv, mapreduce, configuration, xml, serialization
• Assignee: Unassigned Assign to me
• Reporter: Harsh J
• Created: 11/Mar/11 04:20
• Updated: 20/Nov/15 12:41
• Resolved: 11/Mar/11 04:45

Description

The current importtsv functionality will not work if one passes a non-XML representable character as the separator character (say, an escape character - \u001b, fairly common in use).

-Dimporttsv.separator=$'\x1b' # This param fails the submitter when serialized.

While this is a limitation with the Configuration class's being serialized as an XML, it can be circumvented by applying a suitable encoding that makes a string XML-compatible.