smiklosovic commented on code in PR #4523:
URL: https://github.com/apache/cassandra/pull/4523#discussion_r2620306764
##########
src/java/org/apache/cassandra/db/compression/CompressionDictionaryManagerMBean.java:
##########
@@ -32,12 +32,32 @@ public interface CompressionDictionaryManagerMBean
* If no SSTables are available, automatically flushes the memtable first.
* This operation runs synchronously and blocks until training completes.
*
- * @param force force the dictionary training even if there are not enough
samples;
- * otherwise, dictionary training won't start if the trainer
is not ready
+ * @param force force the dictionary training even if
there are not enough samples;
+ * otherwise, dictionary training won't
start if the trainer is not ready
+ * @param maxDictionarySize maximum dictionary size to create, if
not set, CQL parameter value is used
+ * @param maxTotalSampleSize maximum total sample size, if not
set, CQL parameter value is used
* @throws UnsupportedOperationException if table doesn't support
dictionary compression
* @throws IllegalStateException if no SSTables available after
flush
*/
- void train(boolean force);
+ void train(boolean force, String maxDictionarySize, String
maxTotalSampleSize);
Review Comment:
It would be probably better if this was as map so we will not need to change
this everytime some new traning parameter is introduced. If there is a table
backed by a dictionary compressor which accepts yet more parameters (hard to
imagine that now though) it would be way more comfortable.
We also want to add "sampling strategies" under
[CASSANDRA-20935](https://issues.apache.org/jira/browse/CASSANDRA-20935) so I
can imagine that we would put more stuff into this.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
