[
https://issues.apache.org/jira/browse/DRILL-1908?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14261369#comment-14261369
]
Deneche A. Hakim commented on DRILL-1908:
-----------------------------------------
Once a batch is ready to be processed, here is what we should do:
# we allocate the aggregate vectors in the container
# if we have a pending window frame goto 3
# we compute the limits of the window frame while computing the value of the
aggregate function (over the saved batches)
# we write in the container, the aggregate value for all rows of the window
frame. We keep track of the last row written in the container
# if the window frame contains rows from more than one saved batch, we save the
window as pending
otherwise, we clear the window frame
# if we computed the aggregate value for all rows of the container, we are
ready to send the container downstream, transfer the rest of the vectors from
the first saved batch before clearing it.
otherwise, goto 1
> new window function implementation
> ----------------------------------
>
> Key: DRILL-1908
> URL: https://issues.apache.org/jira/browse/DRILL-1908
> Project: Apache Drill
> Issue Type: Improvement
> Components: Execution - Operators
> Reporter: Deneche A. Hakim
> Assignee: Deneche A. Hakim
> Priority: Critical
>
> In order to fix DRILL-1487 a complete rewrite of the
> StreamingWindowFrameRecordBatch may be needed. The purpose of this issue is
> to report my progress and share my thoughts with the community in order to
> get a proper implementation
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
