You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@druid.apache.org by GitBox <gi...@apache.org> on 2022/01/07 05:51:43 UTC

[GitHub] [druid] imply-cheddar opened a new pull request #12130: Add a small LRU cache and use utf8 bytes in ArrayOfDoubles

imply-cheddar opened a new pull request #12130:
URL: https://github.com/apache/druid/pull/12130


   The performance of the ArrayOfDoubles Sketch operations tend to be dominated by stringifying the column used for the sketch.  This change adds a small 10-item LRU cache that allows us to skip this stringification when data is sorted by the key being used *and* also adjusts to use the `lookupUtf8Bytes` method instead to get the value when it is possible.  This has generated a significant speedup in the wild and doesn't change functionality.
   
   <hr>
   
   <!-- Check the items by putting "x" in the brackets for the done things. Not all of these items apply to every PR. Remove the items which are not done or not relevant to the PR. None of the items from the checklist below are strictly necessary, but it would be very helpful if you at least self-review the PR. -->
   
   This PR has:
   - [x] been self-reviewed.
      - [ ] using the [concurrency checklist](https://github.com/apache/druid/blob/master/dev/code-review/concurrency.md) (Remove this item if the PR doesn't have any relation to concurrency.)
   - [ ] added documentation for new or modified features or behaviors.
   - [ ] added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
   - [ ] added or updated version, license, or notice information in [licenses.yaml](https://github.com/apache/druid/blob/master/dev/license.md)
   - [ ] added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
   - [ ] added unit tests or modified existing tests to cover new code paths, ensuring the threshold for [code coverage](https://github.com/apache/druid/blob/master/dev/code-review/code-coverage.md) is met.
   - [ ] added integration tests.
   - [ ] been tested in a test Druid cluster.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org


[GitHub] [druid] clintropolis commented on a change in pull request #12130: Add a small LRU cache and use utf8 bytes in ArrayOfDoubles

Posted by GitBox <gi...@apache.org>.
clintropolis commented on a change in pull request #12130:
URL: https://github.com/apache/druid/pull/12130#discussion_r781899862



##########
File path: extensions-core/datasketches/src/main/java/org/apache/druid/query/aggregation/datasketches/tuple/ArrayOfDoublesSketchBuildAggregator.java
##########
@@ -55,7 +69,10 @@ public ArrayOfDoublesSketchBuildAggregator(
     this.valueSelectors = valueSelectors.toArray(new BaseDoubleColumnValueSelector[0]);
     values = new double[valueSelectors.size()];
     sketch = new ArrayOfDoublesUpdatableSketchBuilder().setNominalEntries(nominalEntries)
-        .setNumberOfValues(valueSelectors.size()).build();
+                                                       .setNumberOfValues(valueSelectors.size()).build();
+
+    this.canCacheById = this.keySelector.nameLookupPossibleInAdvance();

Review comment:
       I think it might be more correct to use `ColumnCapabilities.areDictionaryValuesUnique` instead of `DimensionDictionarySelector.nameLookupPossibleInAdvance`, because the latter does not necessarily require that dictionary ids uniquely identify lookup values, just that it can be done without the cursor being on the current row that provided the dictionary ids. `IndexedTable` selectors for example use the row number as the dictionary id iirc. The capabilities could be retrieved from the column selector factory available in the factorize methods to pass in I think.




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org


[GitHub] [druid] clintropolis merged pull request #12130: Add a small LRU cache and use utf8 bytes in ArrayOfDoubles

Posted by GitBox <gi...@apache.org>.
clintropolis merged pull request #12130:
URL: https://github.com/apache/druid/pull/12130


   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org