[
https://issues.apache.org/jira/browse/KYLIN-2574?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
hongbin ma updated KYLIN-2574:
------------------------------
Request participants: (was: )
Description:
currently RawQueryLastHacker make the raw query group by columns existing in
query (if (tupleInfo.hasColumn(col))). The approach would fail to leverage
limit push down if the existing columns are not a "prefix" of row
keys.(org.apache.kylin.storage.gtrecord.GTCubeStorageQueryBase#enableStorageLimitIfPossible)
On the other hand, a large portion of the raw queries are random queries like
"select * from fact " or "select * from fact inner join lookup where year
=2000" . Keeping these queries return fast is important to impress users
was:currently RawQueryLastHacker make the raw query group by columns existing
in query (if (tupleInfo.hasColumn(col))). The approach would fail to leverage
limit push down if the existing columns are not a "prefix" of row
keys.(org.apache.kylin.storage.gtrecord.GTCubeStorageQueryBase#enableStorageLimitIfPossible)
> RawQueryLastHacker should group by all possible dimensions
> ----------------------------------------------------------
>
> Key: KYLIN-2574
> URL: https://issues.apache.org/jira/browse/KYLIN-2574
> Project: Kylin
> Issue Type: Bug
> Reporter: hongbin ma
> Assignee: hongbin ma
>
> currently RawQueryLastHacker make the raw query group by columns existing in
> query (if (tupleInfo.hasColumn(col))). The approach would fail to leverage
> limit push down if the existing columns are not a "prefix" of row
> keys.(org.apache.kylin.storage.gtrecord.GTCubeStorageQueryBase#enableStorageLimitIfPossible)
> On the other hand, a large portion of the raw queries are random queries like
> "select * from fact " or "select * from fact inner join lookup where year
> =2000" . Keeping these queries return fast is important to impress users
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
