emkornfield commented on code in PR #197:
URL: https://github.com/apache/parquet-format/pull/197#discussion_r1320143220
##########
src/main/thrift/parquet.thrift:
##########
@@ -977,6 +1038,25 @@ struct ColumnIndex {
/** A list containing the number of null values for each page **/
5: optional list<i64> null_counts
+
+ /**
+ * Contains repetition level histograms for more details) for each page
+ * concatenated together. The repetition_level_histogram field on
+ * SizeStatistics contains more details.
+ *
+ * When present the length should always be (number of pages *
+ * (max_repetition_level + 1)) elements in size.
+ *
+ * Element 0 is the first element of the histogram for the first page.
+ * Element (max_repetition_level + 1) is the first element of the histogram
+ * for the second page.
Review Comment:
I'll also point out that if we are very concerned about eacking out the most
efficiency for reads then probably the best organization here is having levels
grouped by pages, and having each entry be the cumulative sum:
`[<page 0, histogram[0]>, <page 1, histogram[0]>, <page 0, histogram[0] +
histogram[1]>, <page 1, histogram[0] + histogram[1]> ]`.
this allows comparing, all not null doing a vectorized subtraction and
checking for positive elements of the results. I'm happy to change this if we
don't think this is too much complexity.
--
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]
