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

Add Null Mask to Prefix and Suffix Iters #6312

Closed
wants to merge 1 commit into from
Closed

Add Null Mask to Prefix and Suffix Iters #6312

Show file tree
Hide file tree
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
30 changes: 14 additions & 16 deletions arrow-array/src/array/byte_view_array.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -318,49 +318,47 @@ impl<T: ByteViewType + ?Sized> GenericByteViewArray<T> {
}

/// Returns an iterator over the prefix bytes of this array with respect to the prefix length.
/// If the prefix length is larger than the string length, it will return the empty slice.
pub fn prefix_bytes_iter(&self, prefix_len: usize) -> impl Iterator<Item = &[u8]> {
/// If the prefix length is larger than the string length, it will return None.
///
/// Note for null views, it returns None.
pub fn prefix_bytes_iter(&self, prefix_len: usize) -> impl Iterator<Item = Option<&[u8]>> {
self.views().into_iter().map(move |v| {
let len = (*v as u32) as usize;

if len < prefix_len {
return &[] as &[u8];
}

if prefix_len <= 4 || len <= 12 {
unsafe { StringViewArray::inline_value(v, prefix_len) }
None
Copy link
Contributor

Choose a reason for hiding this comment

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

What happens to performance if you don't check the null mask, but still return Option

Copy link
Contributor Author

Choose a reason for hiding this comment

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

is_null is cached null values which might faster.

Copy link
Contributor

Choose a reason for hiding this comment

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

I was suggesting simply not checking the null mask at all

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Sorry, I didn't get... The latest commit has removed the is_null check...

Copy link
Contributor Author

@xinlifoobar xinlifoobar Aug 28, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

This comment still points to the old code... I have squash'ed everything in one commit and please see the changes.

} else if prefix_len <= 4 || len <= 12 {
Some(unsafe { StringViewArray::inline_value(v, prefix_len) })
} else {
let view = ByteView::from(*v);
let data = unsafe {
self.data_buffers()
.get_unchecked(view.buffer_index as usize)
};
let offset = view.offset as usize;
unsafe { data.get_unchecked(offset..offset + prefix_len) }
Some(unsafe { data.get_unchecked(offset..offset + prefix_len) })
}
})
}

/// Returns an iterator over the suffix bytes of this array with respect to the suffix length.
/// If the suffix length is larger than the string length, it will return the empty slice.
pub fn suffix_bytes_iter(&self, suffix_len: usize) -> impl Iterator<Item = &[u8]> {
/// If the suffix length is larger than the string length, it will return None.
pub fn suffix_bytes_iter(&self, suffix_len: usize) -> impl Iterator<Item = Option<&[u8]>> {
self.views().into_iter().map(move |v| {
let len = (*v as u32) as usize;

if len < suffix_len {
return &[] as &[u8];
}

if len <= 12 {
unsafe { &StringViewArray::inline_value(v, len)[len - suffix_len..] }
None
} else if len <= 12 {
Some(unsafe { &StringViewArray::inline_value(v, len)[len - suffix_len..] })
} else {
let view = ByteView::from(*v);
let data = unsafe {
self.data_buffers()
.get_unchecked(view.buffer_index as usize)
};
let offset = view.offset as usize;
unsafe { data.get_unchecked(offset + len - suffix_len..offset + len) }
Some(unsafe { data.get_unchecked(offset + len - suffix_len..offset + len) })
}
})
}
Expand Down
16 changes: 12 additions & 4 deletions arrow-string/src/predicate.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,11 @@ impl<'a> Predicate<'a> {
string_view_array
.prefix_bytes_iter(v.len())
.map(|haystack| {
equals_bytes(haystack, v.as_bytes(), equals_kernel) != negate
equals_bytes(
haystack.unwrap_or_default(),
v.as_bytes(),
equals_kernel,
) != negate
})
.collect::<Vec<_>>(),
)
Expand All @@ -151,7 +155,7 @@ impl<'a> Predicate<'a> {
.prefix_bytes_iter(v.len())
.map(|haystack| {
equals_bytes(
haystack,
haystack.unwrap_or_default(),
v.as_bytes(),
equals_ignore_ascii_case_kernel,
) != negate
Expand All @@ -170,7 +174,11 @@ impl<'a> Predicate<'a> {
string_view_array
.suffix_bytes_iter(v.len())
.map(|haystack| {
equals_bytes(haystack, v.as_bytes(), equals_kernel) != negate
equals_bytes(
haystack.unwrap_or_default(),
v.as_bytes(),
equals_kernel,
) != negate
})
.collect::<Vec<_>>(),
)
Expand All @@ -187,7 +195,7 @@ impl<'a> Predicate<'a> {
.suffix_bytes_iter(v.len())
.map(|haystack| {
equals_bytes(
haystack,
haystack.unwrap_or_default(),
v.as_bytes(),
equals_ignore_ascii_case_kernel,
) != negate
Expand Down
Loading