You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cloudberry.apache.org by "leborchuk (via GitHub)" <gi...@apache.org> on 2025/07/14 11:46:50 UTC
Re: [PR] ORCA: allow different strategy control the redistribute key below aggregate [cloudberry]
leborchuk commented on code in PR #1192:
URL: https://github.com/apache/cloudberry/pull/1192#discussion_r2204698041
##########
src/backend/utils/misc/guc_gp.c:
##########
@@ -4488,6 +4489,17 @@ struct config_int ConfigureNamesInt_gp[] =
NULL, NULL, NULL
},
+ {
+ {"optimizer_agg_pds_strategy", PGC_USERSET, DEVELOPER_OPTIONS,
+ gettext_noop("Set the strategy of agg required distribution."),
+ NULL,
+ GUC_NOT_IN_SAMPLE
+ },
+ &optimizer_agg_pds_strategy,
+ OPTIMIZER_AGG_PDS_ALL_KEY, OPTIMIZER_AGG_PDS_ALL_KEY, OPTIMIZER_AGG_PDS_MINIMAL_LEN_KEY,
Review Comment:
Sorry, but if I understand correctly, the goal we are pursuing is to exclude long strings from hash calculations. For example, if we have 10 columns and 5th column is a long string, it would be better not to hash it, and hash all the remaining columns in order to achieve uniform distribution of data across segments.
From this logic, perhaps we could add a GUC with the maximum data type size for hash to exclude all long strings. If all columns contain long strings, then let's select the one with the minimum size.
--
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@cloudberry.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cloudberry.apache.org
For additional commands, e-mail: commits-help@cloudberry.apache.org