[ 
https://issues.apache.org/jira/browse/HIVE-7048?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Swarnim Kulkarni updated HIVE-7048:
-----------------------------------

    Attachment: HIVE-7048.1.patch.txt

[~xuefuz] Thinking further on this, my thought was that it's probably better to 
leave the setupFilter method as a no-op and let the consumer override it 
according to their construction of the composite key implementation. I thought 
that is probably better than setting up a wrong filter and getting wrong 
results.

Another approach was to probably use a PrefixFilter and filter out on the key 
prefixes. It would work but probably won't buy us much in terms of performance 
unless we decide to filter out on the specific pushdown value as well (for 
which we need to know the construction of the keys to get the appropriate bytes 
for the prefix).

Let me know your thoughts on this or if you have a better approach. Thanks!

> CompositeKeyHBaseFactory should not use FamilyFilter
> ----------------------------------------------------
>
>                 Key: HIVE-7048
>                 URL: https://issues.apache.org/jira/browse/HIVE-7048
>             Project: Hive
>          Issue Type: Improvement
>          Components: HBase Handler
>            Reporter: Swarnim Kulkarni
>            Assignee: Swarnim Kulkarni
>            Priority: Blocker
>         Attachments: HIVE-7048.1.patch.txt
>
>
> HIVE-6411 introduced a more generic way to provide composite key 
> implementations via custom factory implementations. However it seems like the 
> CompositeHBaseKeyFactory implementation uses a FamilyFilter for row key scans 
> which doesn't seem appropriate. This should be investigated further and if 
> possible replaced with a RowRangeScanFilter.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to