[FLUME-1945] HBase Serializer allow key from regular expression group

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.3.1
• Fix Version/s: 1.4.0

Release Note

Allow users to chose a regular expression group as hbase row key. To use this feature users need to:

1. Specify the row key index.
2. Specify ROW_KEY as the column name for the corresponding index

For example:
a1.sinks.k1.serializer.rowKeyIndex = 1
a1.sinks.k1.serializer.colNames = host,ROW_KEY,timestamp

• Assignee: Sravya Tirukkovalur
• Reporter: Brock Noland
• Created: 12/Mar/13 11:59
• Updated: 04/Jun/13 20:53
• Resolved: 04/Jun/13 04:46

Description

Back when we created the the HBase serializer interface we decided that most people would have to implement their own since they knew their data and generically mapping an event to columns and keys was likely going to be difficult. I have observed that something like the following: https://github.com/brockn/flume-hbase-serializer

would be useful to users. Basically it's the regex serializer where a single value from the regex match can be used as the key. It might be possible to modify the current regex serializer to behave in this way so as to limit the number of serializers.

Issue Links

• links to: Review board