question Re: hbase snapshot accessing older version of cell in Archives of Support Questions (Read Only)

hbase snapshot accessing older version of cell

aervits — Wed, 28 Sep 2016 00:05:00 GMT

I don't see the levels of detail in hbase reference guide whether cell version can be accessed via hbase snapshot. Use case is I'd like to access an older version of cell via hive over hbase snapshot or via spark, is that information retained in snapshot and do I need HBase to access it or this info is retained in hfile and I can access it with anything other than hbase?

Re: hbase snapshot accessing older version of cell

rgelhausen — Wed, 28 Sep 2016 00:07:20 GMT

@Artem Ervits

See @Dan Zaratsian's examples reading cell versions and timestamps from a snapshot here.

Re: hbase snapshot accessing older version of cell

aervits — Wed, 28 Sep 2016 00:19:34 GMT

@Randy Gelhausen I don't see any reference to HBase cell versions here

Re: hbase snapshot accessing older version of cell

rgelhausen — Wed, 28 Sep 2016 00:22:43 GMT

He filters all cells by timestamp here. If I understand correctly, without the filter, the DataFrame would expose all cell versions that exist in the snapshot.

Re: hbase snapshot accessing older version of cell

aervits — Wed, 28 Sep 2016 00:24:58 GMT

I'll give it a whirl and let you know, until then you get +1

Re: hbase snapshot accessing older version of cell

aervits — Wed, 28 Sep 2016 00:45:04 GMT

@Randy Gelhausen confirmed with @Devaraj Das snapshot captures all info including version information for completeness, here's an example http://stackoverflow.com/questions/28334940/how-to-get-all-versions-of-an-hbase-cell-in-a-spark-newapihadooprdd just need to apply to snapshot rather than table

Re: hbase snapshot accessing older version of cell

dzaratsian — Fri, 30 Sep 2016 02:00:01 GMT

@Artem Ervits, @Randy Gelhausen

Based on your discussion, I added a few additional lines to the github repo mentioned above. The code will now return older versions of a cell. The maximum number of snapshot versions to fetch is specified in the props file (here).