I would go on IRC - irc.freenode.net - and go to the #datastax-brisk and ask driftx (Brandon) if he has any suggestions if you're already using pig 0.9 - that should be the version that has the nested subcolumns fix on the pig side of things.
On Aug 17, 2011, at 3:25 PM, Fabio Souto wrote: > Hi Jeremy, > > Well I think the version it's not the problem, I'm using cassandra 0.8.3 and > pig 0.9, also I tried with brisk beta 2(pig 0.8.3). I comment on the ticket > because it was not clear if it's solved or not, so maybe we can close it. > > Thanks for the response > > Fabio > > On 17/08/2011, at 22:14, Jeremy Hanna wrote: > >> Hi Fabio, >> >> I'm not sure if super columns are fully supported right now in >> CassandraStorage. Brandon (who I CCed) would know for sure. That and I >> thought the pig bug that made it impossible to get to nested data structures >> has been resolved - the ticket you commented on today I think was a >> duplicate of another bug that has been resolved. >> >> What version of pig and what version of cassandra are you using? >> >> Jeremy >> >> On Aug 17, 2011, at 10:09 AM, Fabio Souto wrote: >> >>> Hi, >>> >>> I have some metrics stored on a Cassandra supercolumn and the subcolumns >>> are the timestamps of each metric, I'm loading the metrics in pig with this >>> line: >>> >>> all_metrics = LOAD 'cassandra://keyspace/metrics' USING CassandraStorage() >>> AS (metric_key, metrics_bag: bag {metric: tuple(timestamp, columns: bag >>> {record: tuple(name:chararray, value:chararray)})}); >>> >>> I just want to access the timestamp subcolumn to get the most recent value >>> (using max), I try to use: >>> >>> metric_status = FOREACH all_metrics GENERATE metric_key, >>> metrics_bag.timestamp; >>> dump metric_status; >>> >>> but I'm getting empty values, like this: >>> >>> (key1,{(),(),()}) >>> (key2,{(),()}) >>> ... >>> >>> >>> >> >