Skip to content

fix: array to array cast #2897

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

Open
manuzhang wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: array to array cast #2897

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
23 changes: 15 additions & 8 deletions native/spark-expr/src/conversion_funcs/cast.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,6 @@ use arrow::array::builder::StringBuilder;
use arrow::array::{
BinaryBuilder, DictionaryArray, GenericByteArray, ListArray, MapArray, StringArray, StructArray,
};
use arrow::compute::can_cast_types;
use arrow::datatypes::{ArrowDictionaryKeyType, ArrowNativeType, DataType, Schema};
use arrow::datatypes::{Field, Fields, GenericBinaryType};
use arrow::error::ArrowError;
Expand Down Expand Up @@ -294,6 +293,7 @@ pub(crate) fn cast_array(
};

let cast_result = match (&from_type, to_type) {
(Null, _) => Ok(cast_with_options(&array, to_type, &native_cast_options)?),
(Utf8, Boolean) => spark_cast_utf8_to_boolean::<i32>(&array, eval_mode),
(LargeUtf8, Boolean) => spark_cast_utf8_to_boolean::<i64>(&array, eval_mode),
(Utf8, Timestamp(_, _)) => {
Expand Down Expand Up @@ -366,8 +366,18 @@ pub(crate) fn cast_array(
cast_options,
)?),
(List(_), Utf8) => Ok(cast_array_to_string(array.as_list(), cast_options)?),
(List(_), List(_)) if can_cast_types(&from_type, to_type) => {
Ok(cast_with_options(&array, to_type, &CAST_OPTIONS)?)
(List(_), List(to)) => {
let list_array = array.as_list::<i32>();
Ok(Arc::new(ListArray::new(
Arc::clone(to),
list_array.offsets().clone(),
cast_array(
Arc::clone(list_array.values()),
to.data_type(),
cast_options,
)?,
list_array.nulls().cloned(),
)) as ArrayRef)
}
(Map(_, _), Map(_, _)) => Ok(cast_map_to_map(&array, &from_type, to_type, cast_options)?),
(UInt8 | UInt16 | UInt32 | UInt64, Int8 | Int16 | Int32 | Int64)
Expand Down Expand Up @@ -803,7 +813,8 @@ fn cast_binary_formatter(value: &[u8]) -> String {
#[cfg(test)]
mod tests {
use super::*;
use arrow::array::StringArray;
use arrow::array::{ListArray, StringArray};
use arrow::buffer::OffsetBuffer;
use arrow::datatypes::TimestampMicrosecondType;
use arrow::datatypes::{Field, Fields};
#[test]
Expand Down Expand Up @@ -929,8 +940,6 @@ mod tests {

#[test]
fn test_cast_string_array_to_string() {
use arrow::array::ListArray;
use arrow::buffer::OffsetBuffer;
let values_array =
StringArray::from(vec![Some("a"), Some("b"), Some("c"), Some("a"), None, None]);
let offsets_buffer = OffsetBuffer::<i32>::new(vec![0, 3, 5, 6, 6].into());
Expand All @@ -955,8 +964,6 @@ mod tests {

#[test]
fn test_cast_i32_array_to_string() {
use arrow::array::ListArray;
use arrow::buffer::OffsetBuffer;
let values_array = Int32Array::from(vec![Some(1), Some(2), Some(3), Some(1), None, None]);
let offsets_buffer = OffsetBuffer::<i32>::new(vec![0, 3, 5, 6, 6].into());
let item_field = Arc::new(Field::new("item", DataType::Int32, true));
Expand Down
28 changes: 26 additions & 2 deletions spark/src/test/scala/org/apache/comet/CometCastSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1266,6 +1266,30 @@ class CometCastSuite extends CometTestBase with AdaptiveSparkPlanHelper {
}
}

test("cast ArrayType to ArrayType") {
val types = Seq(
BooleanType,
StringType,
ByteType,
IntegerType,
LongType,
ShortType,
DecimalType(10, 2),
DecimalType(38, 18))
for (fromType <- types) {
for (toType <- types) {
if (fromType != toType &&
!tags
.get(s"cast $fromType to $toType")
.exists(s => s.contains("org.scalatest.Ignore")) &&
Cast.canCast(fromType, toType) &&
CometCast.isSupported(fromType, toType, None, CometEvalMode.LEGACY) == Compatible()) {
castTest(generateArrays(100, fromType), ArrayType(toType))
}
}
}
}

private def generateFloats(): DataFrame = {
withNulls(gen.generateFloats(dataSize)).toDF("a")
}
Expand Down Expand Up @@ -1294,10 +1318,10 @@ class CometCastSuite extends CometTestBase with AdaptiveSparkPlanHelper {
withNulls(gen.generateLongs(dataSize)).toDF("a")
}

private def generateArrays(rowSize: Int, elementType: DataType): DataFrame = {
private def generateArrays(rowNum: Int, elementType: DataType): DataFrame = {
Copy link
Contributor

@coderfender coderfender Jan 10, 2026

Choose a reason for hiding this comment

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

nit : might be an unintended change ?

Copy link
Member Author

@manuzhang manuzhang Jan 11, 2026

Choose a reason for hiding this comment

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

This is intended since rowNum (meaning the number of rows) makes more sense to me.

import scala.collection.JavaConverters._
val schema = StructType(Seq(StructField("a", ArrayType(elementType), true)))
spark.createDataFrame(gen.generateRows(rowSize, schema).asJava, schema)
spark.createDataFrame(gen.generateRows(rowNum, schema).asJava, schema)
}

// https://github.com/apache/datafusion-comet/issues/2038
Expand Down
Loading