You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Sergey Nuyanzin (JIRA)" <ji...@apache.org> on 2018/05/25 08:15:00 UTC
[jira] [Comment Edited] (FLINK-8650) Add tests and documentation
for WINDOW clause
[ https://issues.apache.org/jira/browse/FLINK-8650?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16490391#comment-16490391 ]
Sergey Nuyanzin edited comment on FLINK-8650 at 5/25/18 8:14 AM:
-----------------------------------------------------------------
Hello [~twalthr] I would take this issue if you do not mind
a question for clarification: am I right that a good point to start is _org.apache.flink.table.runtime.stream.sql.OverWindowITCase_
and check the same queries with the same data but with window clause as you mentioned e.g.
now in _org.apache.flink.table.runtime.stream.sql.OverWindowITCase#testProcTimeBoundedPartitionedRowsOver_ there is
{code:sql}
SELECT
a,
SUM(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW),
MIN(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
FROM MyTable
{code}
which could be checked against
{code:sql}
SELECT
a,
SUM(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW),
MIN(c) OVER w
FROM MyTable WINDOW w AS (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
{code}
{code:sql}
SELECT
a,
SUM(c) OVER w,
MIN(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
FROM MyTable WINDOW w AS (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
{code}
{code:sql}
SELECT
a,
SUM(c) OVER w,
MIN(c) OVER w
FROM MyTable WINDOW w AS (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
{code}
I guess one thing is it will be nice to do it not manually via copy paste but in a more intelligent approach
was (Author: sergey nuyanzin):
Hello [~twalthr] I would take this issue if you do not mind
a question for clarification: am I right that a good point to start is _org.apache.flink.table.runtime.stream.sql.OverWindowITCase_
and check the same queries with the same data but with window clause as you mentioned e.g.
now in _org.apache.flink.table.runtime.stream.sql.OverWindowITCase#testProcTimeBoundedPartitionedRowsOver_ there is
{code:sql}
SELECT
a,
SUM(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW),
MIN(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
FROM MyTable
{code}
which could be checked against
{code:sql}
SELECT
a,
SUM(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW),
MIN(c) OVER w
FROM MyTable WINDOW w AS OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
{code}
{code:sql}
SELECT
a,
SUM(c) OVER w,
MIN(c) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
FROM MyTable WINDOW w AS OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
{code}
{code:sql}
SELECT
a,
SUM(c) OVER w,
MIN(c) OVER w
FROM MyTable WINDOW w AS OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
{code}
I guess one thing is it will be nice to do it not manually via copy paste but in a more intelligent approach
> Add tests and documentation for WINDOW clause
> ---------------------------------------------
>
> Key: FLINK-8650
> URL: https://issues.apache.org/jira/browse/FLINK-8650
> Project: Flink
> Issue Type: Improvement
> Components: Table API & SQL
> Reporter: Timo Walther
> Assignee: Sergey Nuyanzin
> Priority: Major
>
> We support queries with a {{WINDOW}} clause like:
> {code}
> SELECT a, SUM(c) OVER w, MIN(c) OVER w FROM MyTable WINDOW w AS (PARTITION BY a ORDER BY proctime ROWS BETWEEN 4 PRECEDING AND CURRENT ROW)
> {code}
> But this is neither documented nor tested.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)