[jira] Commented: (HBASE-725) Maximum ColumnKey Lookup

["Jim Kellerman (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-725?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12611205#action_12611205 ] 

Jim Kellerman commented on HBASE-725:
-------------------------------------

By time series relative column keys, are you talking about the row timestamp, or are you storing the data as family:nnnnnnnnn where nnnnnnnnn is the timestamp?

If the latter, there is no good way to search for the "last" one, because the part of the column key after the ':' can be anything. If you are storing time series data I would recommend using the row timestamp because each cell is addressed by row/family:member/timestamp

> Maximum ColumnKey Lookup
> ------------------------
>
>                 Key: HBASE-725
>                 URL: https://issues.apache.org/jira/browse/HBASE-725
>             Project: Hadoop HBase
>          Issue Type: New Feature
>            Reporter: Daniel Blaisdell
>            Priority: Minor
>             Fix For: 0.2.0
>
>
> When working with time series data, relative columnKeys 'indexes' are routinely shared between the columnFamilies of a given row. To acquire the maximum key and increment it for the next columnKey, you must first get all the keys. A quick lookup of the last columnKey would benefit as the column size grows over time. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.