Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Minor: Refactor array_union function to use a generic union_arrays function #8381

Merged
merged 1 commit into from
Dec 1, 2023
Merged

Minor: Refactor array_union function to use a generic union_arrays function #8381

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 21 additions & 20 deletions datafusion/physical-expr/src/array_expressions.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1515,32 +1515,33 @@ pub fn array_union(args: &[ArrayRef]) -> Result<ArrayRef> {
}
let array1 = &args[0];
let array2 = &args[1];

fn union_arrays<O: OffsetSizeTrait>(
array1: &ArrayRef,
array2: &ArrayRef,
l_field_ref: &Arc<Field>,
r_field_ref: &Arc<Field>,
) -> Result<ArrayRef> {
match (l_field_ref.data_type(), r_field_ref.data_type()) {
(DataType::Null, _) => Ok(array2.clone()),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Awesome @Weijun-H
Makes the code cleaner

Its not a problem of this PR but I'm thinking why we need to clone arrays in case of nulls.... 🤔

Copy link
Member Author

@Weijun-H Weijun-H Dec 1, 2023 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If array1 is Null, return array2 directly because it is a union function. @comphead

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nvm, thanks @Weijun-H its cloning a reference not the underlying data

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the whole handling of Nulls in array functions is not quite correct (DataType::Null) should not be passed in. I believe @jayzhan211 is working on this

(_, DataType::Null) => Ok(array1.clone()),
(_, _) => {
let list1 = array1.as_list::<O>();
let list2 = array2.as_list::<O>();
let result = union_generic_lists::<O>(list1, list2, l_field_ref)?;
Ok(Arc::new(result))
}
}
}

match (array1.data_type(), array2.data_type()) {
(DataType::Null, _) => Ok(array2.clone()),
(_, DataType::Null) => Ok(array1.clone()),
(DataType::List(l_field_ref), DataType::List(r_field_ref)) => {
match (l_field_ref.data_type(), r_field_ref.data_type()) {
(DataType::Null, _) => Ok(array2.clone()),
(_, DataType::Null) => Ok(array1.clone()),
(_, _) => {
let list1 = array1.as_list::<i32>();
let list2 = array2.as_list::<i32>();
let result = union_generic_lists::<i32>(list1, list2, l_field_ref)?;
Ok(Arc::new(result))
}
}
union_arrays::<i32>(array1, array2, l_field_ref, r_field_ref)
}
(DataType::LargeList(l_field_ref), DataType::LargeList(r_field_ref)) => {
match (l_field_ref.data_type(), r_field_ref.data_type()) {
(DataType::Null, _) => Ok(array2.clone()),
(_, DataType::Null) => Ok(array1.clone()),
(_, _) => {
let list1 = array1.as_list::<i64>();
let list2 = array2.as_list::<i64>();
let result = union_generic_lists::<i64>(list1, list2, l_field_ref)?;
Ok(Arc::new(result))
}
}
union_arrays::<i64>(array1, array2, l_field_ref, r_field_ref)
}
_ => {
internal_err!(
Expand Down
Loading