[
https://issues.apache.org/jira/browse/HIVE-26479?focusedWorklogId=802705&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-802705
]
ASF GitHub Bot logged work on HIVE-26479:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 23/Aug/22 06:15
Start Date: 23/Aug/22 06:15
Worklog Time Spent: 10m
Work Description: SourabhBadhya commented on code in PR #3528:
URL: https://github.com/apache/hive/pull/3528#discussion_r952191010
##########
ql/src/java/org/apache/hadoop/hive/ql/txn/compactor/QueryCompactor.java:
##########
@@ -272,5 +277,23 @@ static void removeFilesForMmTable(HiveConf conf,
AcidDirectory dir) throws IOExc
fs.delete(dead, true);
}
}
+
+ static void overrideConfProps(HiveConf conf, CompactionInfo ci,
Map<String, String> properties) {
+ for (String key : properties.keySet()) {
+ if (key.startsWith(COMPACTOR_PREFIX)) {
+ String property = key.substring(10); // 10 is the length of
"compactor." We only keep the rest.
+ conf.set(property, properties.get(key));
+ }
+ }
+
+ // Give preference to properties coming from compaction
+ // over table properties
+ for (String key : ci.getPropertiesMap().keySet()) {
Review Comment:
This is required for the `ALTER TABLE <tablename> COMPACT 'major' WITH
TBLPROPERTIES OVERWRITE (<property list>)` scenario. Here the properties are
entered into `COMPACTION_QUEUE` and when the table is picked for compaction,
the table properties which are set in `CQ_TBLPROPERTIES` must also be picked.
All this information is stored in CompactionInfo, hence the need to set the
properties as here.
Same behaviour is seen for MR-based compaction as well.
Issue Time Tracking
-------------------
Worklog Id: (was: 802705)
Time Spent: 1h 10m (was: 1h)
> Add ability to set parameters for query-based compaction
> --------------------------------------------------------
>
> Key: HIVE-26479
> URL: https://issues.apache.org/jira/browse/HIVE-26479
> Project: Hive
> Issue Type: Improvement
> Reporter: Sourabh Badhya
> Assignee: Sourabh Badhya
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
> [HIVE-13354|https://issues.apache.org/jira/browse/HIVE-13354] introduced the
> ability to set some parameters for the compaction through table properties,
> like the mapper memory size or compaction thresholds. This could be useful
> for the query-based compaction as well, for example if the insert of the
> query-based compaction is failing, we would have a possibility to tune the
> compaction run directly. First it should be investigated which properties are
> possible and would make sense to set for the query base compaction. Then
> implement this feature for the query-based compaction.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
