How to filter single column in a jtable?

Is there an efficient way to scan and filter by column qualifier in HBase?

• Lets say I have the following Metrics table: row key: <metric_id> Families:   stats_monthly => columns: <YYYYMM>_stat1, <YYYYMM>_stat2,   ...   stats_weekly => columns: <week#>_stat1, ... Is there an efficient way to query for all the stats columns of a certain metric for a given month? (something like all values for "metric1:stats_monthly:201107_*") Or maybe a better (or only?) solution is to split the Metrics table into Metrics_Monthly and Metrics_Daily and move the date part from the column qualifier to the row_key? (row key would be <YYYYMM>_<metric_id> and when querying for a specific month I could scan for all row keys that with <YYYYMM>) Since all writes call will be on the current month wont all the writes going to the same server with this method? Update: is using a ColumnPrefixFilter the right solution? Looking at the code it seems that it enumerates over all the possible KeyValues available which isnt as performant is jumping to the first match and scanning from there as I would expect...

• Answer:

  You may use a ColumnPrefixFilter to filter keys by their columns prefix or you may even implement your own column filter.  Read more about filters here: http://ofps.oreilly.com/titles/9781449396107/clientapisadv.html And here some info about the ColumnPrefixFilter: https://issues.apache.org/jira/browse/HBASE-3684 ColumnPrefixFilter will scan all the columns making the amount of data transfered from Hbase to client smaller, but requiring to scan all the columns. In other words Hbase does not have any index on column names (in contradiction to Cassandra, which has columns stored in treemaps, guaranteeing sorted data and fast submap() functionality). So the solution will be to implement your own index on top of Hbase, resulting for additional maintenance and additional round trips to Hbase cluster. Luckily there is a good implementation already called IHbase , which optimizes the columns access: https://issues.apache.org/jira/browse/HBASE-2037 https://github.com/ykulbak/ihbase/wiki/Getting-Started