You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@arrow.apache.org by GitBox <gi...@apache.org> on 2022/07/13 07:47:57 UTC
[GitHub] [arrow-rs] jhorstmann opened a new issue, #2057: Boolean `take` kernel does not handle null indices correctly
jhorstmann opened a new issue, #2057:
URL: https://github.com/apache/arrow-rs/issues/2057
**Describe the bug**
The boolean `take` kernel does not handle null values in the indices array, instead it always accesses the bitmap and values at the masked off index. Usually this index is the default of 0, but if it is out of bounds this will lead to a panic.
**To Reproduce**
#[test]
fn test_take_bool_nullable_index() {
// indices where the masked invalid elements would be out of bounds
let index_data = ArrayData::try_new(
DataType::Int32,
6,
Some(Buffer::from_iter(vec![false, true, false, true, false, true])),
0,
vec![Buffer::from_iter(vec![99, 0, 999, 1, 9999, 2])],
vec![]
).unwrap();
let index = UInt32Array::from(index_data);
test_take_boolean_arrays(
vec![Some(true), None, Some(false)],
&index,
None,
vec![None, Some(true), None, None, None, Some(false)],
);
}
```
assertion failed: i < (self.bits.len() << 3)
thread 'compute::kernels::take::tests::test_take_bool_nullable_index' panicked at 'assertion failed: i < (self.bits.len() << 3)', arrow/src/bitmap.rs:55:9
```
**Expected behavior**
Null indices should be ignored.
**Additional context**
Could be solved by using `PrimitiveArray::iter` to iterate over indices. If that reduces performance, some of it could probably be regained by #1857
The performance of the `take` kernel might also be improved by taking value and validity bits separately.
--
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: github-unsubscribe@arrow.apache.org.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-rs] tustvold closed issue #2057: Boolean `take` kernel does not handle null indices correctly
Posted by GitBox <gi...@apache.org>.
tustvold closed issue #2057: Boolean `take` kernel does not handle null indices correctly
URL: https://github.com/apache/arrow-rs/issues/2057
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org