Skip to content

Update REVERSE scalar function to support Utf8View #11973

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

Merged
merged 1 commit into from
Aug 15, 2024
Merged

Update REVERSE scalar function to support Utf8View #11973

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
106 changes: 57 additions & 49 deletions datafusion/functions/src/unicode/reverse.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,12 +18,14 @@
use std::any::Any;
use std::sync::Arc;

use arrow::array::{ArrayRef, GenericStringArray, OffsetSizeTrait};
use arrow::array::{
Array, ArrayAccessor, ArrayIter, ArrayRef, AsArray, GenericStringArray,
OffsetSizeTrait,
};
use arrow::datatypes::DataType;

use datafusion_common::cast::as_generic_string_array;
use datafusion_common::{exec_err, Result};
use datafusion_expr::{ColumnarValue, ScalarUDFImpl, Signature, Volatility};
use DataType::{LargeUtf8, Utf8, Utf8View};

use crate::utils::{make_scalar_function, utf8_to_str_type};

Expand All @@ -44,7 +46,7 @@ impl ReverseFunc {
Self {
signature: Signature::uniform(
1,
vec![Utf8, LargeUtf8],
vec![Utf8View, Utf8, LargeUtf8],
Volatility::Immutable,
),
}
Expand All @@ -70,8 +72,8 @@ impl ScalarUDFImpl for ReverseFunc {

fn invoke(&self, args: &[ColumnarValue]) -> Result<ColumnarValue> {
match args[0].data_type() {
DataType::Utf8 => make_scalar_function(reverse::<i32>, vec![])(args),
DataType::LargeUtf8 => make_scalar_function(reverse::<i64>, vec![])(args),
Utf8 | Utf8View => make_scalar_function(reverse::<i32>, vec![])(args),
LargeUtf8 => make_scalar_function(reverse::<i64>, vec![])(args),
other => {
exec_err!("Unsupported data type {other:?} for function reverse")
}
Expand All @@ -83,10 +85,17 @@ impl ScalarUDFImpl for ReverseFunc {
/// reverse('abcde') = 'edcba'
/// The implementation uses UTF-8 code points as characters
pub fn reverse<T: OffsetSizeTrait>(args: &[ArrayRef]) -> Result<ArrayRef> {
let string_array = as_generic_string_array::<T>(&args[0])?;
if args[0].data_type() == &Utf8View {
Copy link
Contributor

Choose a reason for hiding this comment

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

probably we can keep same pattern like in invoke ? it would be more readable

Copy link
Contributor Author

Choose a reason for hiding this comment

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

as in using a match? Sure. It won't be the same match arms though.

reverse_impl::<T, _>(args[0].as_string_view())
} else {
reverse_impl::<T, _>(args[0].as_string::<T>())
}
}

let result = string_array
.iter()
fn reverse_impl<'a, T: OffsetSizeTrait, V: ArrayAccessor<Item = &'a str>>(
string_array: V,
) -> Result<ArrayRef> {
let result = ArrayIter::new(string_array)
.map(|string| string.map(|string: &str| string.chars().rev().collect::<String>()))
.collect::<GenericStringArray<T>>();

Expand All @@ -95,59 +104,58 @@ pub fn reverse<T: OffsetSizeTrait>(args: &[ArrayRef]) -> Result<ArrayRef> {

#[cfg(test)]
mod tests {
use arrow::array::{Array, StringArray};
use arrow::datatypes::DataType::Utf8;
use arrow::array::{Array, LargeStringArray, StringArray};
use arrow::datatypes::DataType::{LargeUtf8, Utf8};

use datafusion_common::{Result, ScalarValue};
use datafusion_expr::{ColumnarValue, ScalarUDFImpl};

use crate::unicode::reverse::ReverseFunc;
use crate::utils::test::test_function;

macro_rules! test_reverse {
Copy link
Contributor

Choose a reason for hiding this comment

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

is it possible to have a function rather than macro?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Could, sure. I'm curious as to why though. This is just the same approach as I used for lpad.

Copy link
Contributor

Choose a reason for hiding this comment

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

yeah this testing pattern is somewhat common in the string functions

I happen to like .slt in general, but I know that is not for everyone

Copy link
Contributor

Choose a reason for hiding this comment

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

I agree, we trying to use macros if other options didn't work. It takes more time when people debug tests to expand the macros and go through them.

if its already used in string tests its okay, perhaps we want to revisit macros and replace them with funcs whenever possible

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Great explanation ,thanks.

($INPUT:expr, $EXPECTED:expr) => {
test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::Utf8($INPUT))],
$EXPECTED,
&str,
Utf8,
StringArray
);

test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::LargeUtf8($INPUT))],
$EXPECTED,
&str,
LargeUtf8,
LargeStringArray
);

test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::Utf8View($INPUT))],
$EXPECTED,
&str,
Utf8,
StringArray
);
};
}

#[test]
fn test_functions() -> Result<()> {
test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::from("abcde"))],
Ok(Some("edcba")),
&str,
Utf8,
StringArray
);
test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::from("loẅks"))],
Ok(Some("sk̈wol")),
&str,
Utf8,
StringArray
);
test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::from("loẅks"))],
Ok(Some("sk̈wol")),
&str,
Utf8,
StringArray
);
test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::Utf8(None))],
Ok(None),
&str,
Utf8,
StringArray
);
test_reverse!(Some("abcde".into()), Ok(Some("edcba")));
test_reverse!(Some("loẅks".into()), Ok(Some("sk̈wol")));
test_reverse!(Some("loẅks".into()), Ok(Some("sk̈wol")));
test_reverse!(None, Ok(None));
#[cfg(not(feature = "unicode_expressions"))]
test_function!(
ReverseFunc::new(),
&[ColumnarValue::Scalar(ScalarValue::from("abcde"))],
test_reverse!(
Some("abcde".into()),
internal_err!(
"function reverse requires compilation with feature flag: unicode_expressions."
),
&str,
Utf8,
StringArray
);

Ok(())
Expand Down
30 changes: 30 additions & 0 deletions datafusion/sqllogictest/test_files/functions.slt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -234,6 +234,16 @@ SELECT reverse('abcde')
----
edcba

query T
SELECT reverse(arrow_cast('abcde', 'LargeUtf8'))
----
edcba

query T
SELECT reverse(arrow_cast('abcde', 'Utf8View'))
----
edcba

query T
SELECT reverse(arrow_cast('abcde', 'Dictionary(Int32, Utf8)'))
----
Expand All @@ -244,11 +254,31 @@ SELECT reverse('loẅks')
----
sk̈wol

query T
SELECT reverse(arrow_cast('loẅks', 'LargeUtf8'))
----
sk̈wol

query T
SELECT reverse(arrow_cast('loẅks', 'Utf8View'))
----
sk̈wol

query T
SELECT reverse(NULL)
----
NULL

query T
SELECT reverse(arrow_cast(NULL, 'LargeUtf8'))
----
NULL

query T
SELECT reverse(arrow_cast(NULL, 'Utf8View'))
----
NULL

query T
SELECT right('abcde', -2)
----
Expand Down
3 changes: 1 addition & 2 deletions datafusion/sqllogictest/test_files/string_view.slt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -884,14 +884,13 @@ logical_plan
03)----TableScan: test projection=[column1_utf8view, column2_utf8view]

## Ensure no casts for REVERSE
## TODO file ticket
query TT
EXPLAIN SELECT
REVERSE(column1_utf8view) as c1
FROM test;
----
logical_plan
01)Projection: reverse(CAST(test.column1_utf8view AS Utf8)) AS c1
01)Projection: reverse(test.column1_utf8view) AS c1
02)--TableScan: test projection=[column1_utf8view]


Expand Down