Uh oh!
There was an error while loading. Please reload this page.
GH-3273: Add scoped chunk level statistics to avoid unbounded output - #3274
Conversation
ArnavBalyan
commented
Aug 22, 2025
cc @wgtmac@shangxinli could you please review thanks! |
1 similar comment
ArnavBalyan
commented
Aug 22, 2025
cc @wgtmac@shangxinli could you please review thanks! |
ArnavBalyan
commented
Aug 27, 2025
cc @gszadovszky could you please help take a look thanks! |
| for (int index = 0, n = rowGroups.size(); index < n; index++) { | ||
| printRowGroupSizeStats(console, index, rowGroups.get(index), schema); | ||
| List<BlockMetaData> blocks = footer.getBlocks(); | ||
| Set<Integer> allowedRowGroups = rowGroups == null ? null : new HashSet<>(rowGroups); |
There was a problem hiding this comment.
nit: If you would create an empty Set instead of initializing with a null, you wouldn't have to check for null in the loop.
There was a problem hiding this comment.
Previous implementation was using map since we were tracking depth, updated to set thanks!
| Set<String> allowedColumns = null; | ||
| if (columns != null && !columns.isEmpty()) { | ||
| allowedColumns = new HashSet<>(columns); | ||
| } |
wgtmac
commented
Aug 27, 2025
Could you please provide the cli output after your improvement? Thanks! |
Here is the CLI output, it's the same as the previous CLI command's output just truncated based on the user's preference: Thanks! |
Uh oh!
There was an error while loading. Please reload this page.
-c, --column: filter by column dot paths (e.g., a.b.c)
-r, --row-group: filter by row-group indexes (0-based)