asolimando commented on code in PR #19957:
URL: https://github.com/apache/datafusion/pull/19957#discussion_r2923969032
##########
datafusion/datasource-parquet/src/metadata.rs:
##########
@@ -541,6 +541,36 @@ fn summarize_min_max_null_counts(
)
.map(|(idx, _)| idx);
+ // Extract distinct counts from row group column statistics
+ accumulators.distinct_counts_array[logical_schema_index] =
+ if let Some(parquet_idx) = parquet_index {
+ let distinct_counts: Vec<u64> = row_groups_metadata
+ .iter()
+ .filter_map(|rg| {
+ rg.columns()
+ .get(parquet_idx)
+ .and_then(|col| col.statistics())
+ .and_then(|stats| stats.distinct_count_opt())
+ })
+ .collect();
+
+ if distinct_counts.is_empty() {
+ Precision::Absent
+ } else if distinct_counts.len() == 1 {
Review Comment:
Thanks for the reminder @gene-bordegaray, I have followed the advise of
@jonathanc-n in faf521161.
Added a `PARTIAL_NDV_THRESHOLD` constant (0.75): if fewer than 75% of row
groups report NDV, return `Absent`. Kept as a constant rather than a config
option since the statistics extraction path has no config access today. It
would require updating a lot of signatures, we can consider that for later.
--
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]
