You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pig.apache.org by "Ted Malaska (JIRA)" <ji...@apache.org> on 2012/11/07 18:09:11 UTC
[jira] [Commented] (PIG-3037) Order by partition by
[ https://issues.apache.org/jira/browse/PIG-3037?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13492513#comment-13492513 ]
Ted Malaska commented on PIG-3037:
----------------------------------
Here is a simple use case
I want to add the max and min price for the day up onto the current record
Ticker|time|price
FB|1|5
FB|2|4
FB|3|6
would output the following
Ticker|time|price|max|min
FB|1|5|5|5
FB|2|4|5|4
FB|3|6|6|4
For every trade of FB in a given day or week.
> Order by partition by
> ---------------------
>
> Key: PIG-3037
> URL: https://issues.apache.org/jira/browse/PIG-3037
> Project: Pig
> Issue Type: Bug
> Reporter: Ted Malaska
> Priority: Minor
>
> Why doesn't PIG support partition by on order by?
> If PIG would then PIG could be used for a mess of windowing functions.
> Is there some reason why we can't add a custom partitioner on a order by operation?
> Is there a work around to do windowing in PIG.
> I understand the I can group and then order with in a group, but what if the items with in the group are super big will I have memory issues? Because I need to order the values with in a group. i.e. stock tickers is the group and they need to be sorted on time.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira