alamb commented on code in PR #2170:
URL: https://github.com/apache/arrow-datafusion/pull/2170#discussion_r845539449
##########
datafusion/core/src/physical_plan/file_format/parquet.rs:
##########
@@ -919,6 +955,73 @@ mod tests {
assert_batches_sorted_eq!(expected, &read);
}
+ #[tokio::test]
+ async fn evolved_schema_filter() {
+ let c1: ArrayRef =
+ Arc::new(StringArray::from(vec![Some("Foo"), None, Some("bar")]));
+
+ let c2: ArrayRef = Arc::new(Int64Array::from(vec![Some(1), Some(2),
None]));
+
+ let c3: ArrayRef = Arc::new(Int8Array::from(vec![Some(10), Some(20),
None]));
+
+ // batch1: c1(string), c2(int64), c3(int8)
+ let batch1 = create_batch(vec![
+ ("c1", c1.clone()),
+ ("c2", c2.clone()),
+ ("c3", c3.clone()),
+ ]);
+
+ // batch2: c3(int8), c2(int64), c1(string)
+ let batch2 = create_batch(vec![("c3", c3), ("c2", c2), ("c1", c1)]);
+
+ let filter = col("c3").eq(lit(0_i8));
+
+ // read/write them files:
+ let read = round_trip_to_parquet(vec![batch1, batch2], None, None,
Some(filter))
+ .await
+ .unwrap();
+
+ // Predicate should prune all row groups
+ assert_eq!(read.len(), 0);
+ }
+
+ #[tokio::test]
+ async fn evolved_schema_disjoint_schema_filter() {
+ let c1: ArrayRef =
+ Arc::new(StringArray::from(vec![Some("Foo"), None, Some("bar")]));
+
+ let c2: ArrayRef = Arc::new(Int64Array::from(vec![Some(1), Some(2),
None]));
Review Comment:
I think the key point is that filtering is *also* applied after the initial
parquet scan / pruning -- the pruning is just a first pass to try and reduce
additional work.
So subsequent `Filter` operations will actually handle filtering out the
columns with `c2 = null`
--
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]
