[PIG-4355] Piggybank: XPath cant handle namespace in xpath, nor can it return more than one match

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.14.0
• Fix Version/s: 0.15.0
• Component/s: piggybank
• Patch Info: Patch Available
• Hadoop Flags: Reviewed
• Assignee: John Cavanaugh
• Reporter: John Cavanaugh
• Created: 11/Dec/14 06:05
• Updated: 07/Jun/15 03:47
• Resolved: 15/Jan/15 19:20

Description

If you pass an xpath that contains a namespace the XPath UDF will always fail to match.

It would be better to either silently remove the namespace or provide a parameter that will remove it.

The reason it is desirable to ignore xpath's with namespaces is that many xml tools when selecting an xpath provide the namespace. It makes cutting & pasting into a pig script painful if you need to manually remove it.

Additionally XPath only returns the first match. It is often desirable to return all matches and allow for a flattening to process multiple records. An XPathAll would be useful to have.

A patch is available as a git pullrequest at
https://github.com/apache/pig/pull/14