[
https://issues.apache.org/jira/browse/ARROW-14035?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17418866#comment-17418866
]
Ian Cook edited comment on ARROW-14035 at 9/22/21, 11:12 PM:
-------------------------------------------------------------
{quote}1. Do we need to compute the same thing of hash_count_distinct but
without using the hash table from the hash group?
{quote}
Yes. [~lidavidm] if I am missing any nuance here, please let me know :)
{quote}Are we going to offer non hash version for all hash_x functions too?
(hash_distinct, hash_count, hash_sum)
{quote}
Yes I think we should aim for that (or nearly that; there might be a few
exceptions where it does not make sense.) Comparing the lists of aggregation
functions and hash (grouped) aggregation functions in
[compute.rst|https://github.com/apache/arrow/blob/master/docs/source/cpp/compute.rst],
they are mostly the same already, with just a few differences. I think this
issue and ARROW-13309 are the most important two additions to bring these two
lists closer to parity.
was (Author: icook):
{quote}1. Do we need to compute the same thing of hash_count_distinct but
without using the hash table from the hash group?
{quote}
Yes
{quote}Are we going to offer non hash version for all hash_x functions too?
(hash_distinct, hash_count, hash_sum)
{quote}
Yes I think we should aim for that (or nearly that; there might be a few
exceptions where it does not make sense.) Comparing the lists of aggregation
functions and hash (grouped) aggregation functions in
[compute.rst|https://github.com/apache/arrow/blob/master/docs/source/cpp/compute.rst],
they are mostly the same already, with just a few differences. I think this
issue and ARROW-13309 are the most important two additions to bring these two
lists closer to parity.
> [C++][Compute] Implement non-hash count_distinct aggregate kernel
> -----------------------------------------------------------------
>
> Key: ARROW-14035
> URL: https://issues.apache.org/jira/browse/ARROW-14035
> Project: Apache Arrow
> Issue Type: New Feature
> Components: C++
> Reporter: Ian Cook
> Assignee: Percy Camilo Triveño Aucahuasi
> Priority: Critical
> Labels: kernel
> Fix For: 6.0.0
>
>
> ARROW-12728 added a {{hash_count_distinct}} hash aggregate kernel, but there
> is no non-hash {{count_distinct}} aggregate kernel.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
