[
https://issues.apache.org/jira/browse/HIVE-29312?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Wechar updated HIVE-29312:
--------------------------
Affects Version/s: (was: 4.1.0)
> Concatenate equality conditions in AND nodes
> --------------------------------------------
>
> Key: HIVE-29312
> URL: https://issues.apache.org/jira/browse/HIVE-29312
> Project: Hive
> Issue Type: Improvement
> Components: Standalone Metastore
> Reporter: Wechar
> Assignee: Wechar
> Priority: Major
>
> Current partition filtert tree is a binary tree with single condition in the
> leaf node, actually the adjacent AND leaf nodes can be combined together to
> match longer index prefix. For example, a table has partition key: a, b, c,
> d, and a filter is {{a=1 and b=1 and c=1}}, when visit as a binary tree
> *AND(AND(a=1, b=1), c=1)*, the filter is:
> {code:sql}
> PART_NAME like 'a=1/%' and PART_NAME like '%/b=2/%' and PART_NAME like
> '%/c=3/%'
> {code}
> If we combine the adjacent AND conditions, the tree will be *MultiAnd(a=1,
> b=2, c=3)*, aand the filte can be:
> {code:bash}
> PART_NAME like 'a=1/b=2/c=3/%'
> {code}
> Obviously, the latter could match a longer index prefix and is more efficient.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
