@@ -669,14 +669,30 @@ impl WriterPropertiesBuilder {
669
669
}
670
670
}
671
671
672
- /// Controls the level of statistics to be computed by the writer
672
+ /// Controls the level of statistics to be computed by the writer and stored in
673
+ /// the parquet file.
674
+ ///
675
+ /// Enabling statistics makes the resulting Parquet file larger and requires
676
+ /// more time to read the parquet footer.
677
+ ///
678
+ /// Statistics can be used to improve query performance by pruning row groups
679
+ /// and pages during query execution if the query engine supports evaluating the
680
+ /// predicate using the statistics.
673
681
#[ derive( Debug , Clone , Copy , Eq , PartialEq ) ]
674
682
pub enum EnabledStatistics {
675
- /// Compute no statistics
683
+ /// Compute no statistics.
676
684
None ,
677
- /// Compute chunk-level statistics but not page-level
685
+ /// Compute column chunk-level statistics but not page-level.
686
+ ///
687
+ /// Setting this option will store one set of statistics for each relevant
688
+ /// column for each row group. The more row groups written, the more
689
+ /// statistics will be stored.
678
690
Chunk ,
679
- /// Compute page-level and chunk-level statistics
691
+ /// Compute page-level and column chunk-level statistics.
692
+ ///
693
+ /// Setting this option will store one set of statistics for each relevant
694
+ /// column for each page and row group. The more row groups and the more
695
+ /// pages written, the more statistics will be stored.
680
696
Page ,
681
697
}
682
698
0 commit comments