[
https://issues.apache.org/jira/browse/HAWQ-1409?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15940520#comment-15940520
]
Kavinder Dhaliwal commented on HAWQ-1409:
-----------------------------------------
Currently the design of this implementation will only support count. So a new
header AGG-TYPE will be sent from HAWQ to PXF with the following possible values
"count"
"unknown"
This simplifies the initial implementation
> HAWQ send additional header to PXF to indicate aggregate function type
> ----------------------------------------------------------------------
>
> Key: HAWQ-1409
> URL: https://issues.apache.org/jira/browse/HAWQ-1409
> Project: Apache HAWQ
> Issue Type: Improvement
> Components: PXF
> Reporter: Kavinder Dhaliwal
> Assignee: Ed Espino
> Fix For: 2.3.0.0-incubating
>
>
> PXF can take advantage of some file formats such as ORC and leverage the
> stats in the metadata. This means that for some simple aggregate functions
> like count, min, max without any complex joins or filters PXF can simply read
> the metadata and avoid reading tuples. In order for PXF to know that a query
> can be completed via ORC metadata HAWQ must indicate to PXF that the query is
> an aggregate query and the type of function
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
