You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "kumarUjjawal (via GitHub)" <gi...@apache.org> on 2026/01/31 14:12:19 UTC
[PR] Refactor i`szero()` and `isnan()` to accept all numeric types [datafusion]
kumarUjjawal opened a new pull request, #20093:
URL: https://github.com/apache/datafusion/pull/20093
## Which issue does this PR close?
<!--
We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closes #123` indicates that this PR will close issue #123.
-->
- Closes #20089
## Rationale for this change
iszero() and isnan() previously accepted “numeric” inputs by implicitly coercing them to Float64, adding unnecessary casts and work for integer/decimal inputs.
<!--
Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes.
-->
## What changes are included in this PR?
- Updated iszero() and isnan() signatures to accept TypeSignatureClass::Numeric without implicit float coercion.
- Refactored iszero() implementation to evaluate zero checks directly for integers, unsigned integers, floats, and decimals.
- Refactored isnan() implementation to compute is_nan only for float types and return false for all other numeric types and reduced non-float array handling to use unary kernels instead of manual iteration.
- Added sqllogictest
<!--
There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR.
-->
## Are these changes tested?
Yes
<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code
If tests are not included in your PR, please explain why (for example, are they covered by existing tests)?
-->
## Are there any user-facing changes?
No
<!--
If there are user-facing changes then we may require documentation to be updated before approving the PR.
-->
<!--
If there are any breaking changes to public APIs, please add the `api change` label.
-->
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Refactor `iszero()` and `isnan()` to accept all numeric types [datafusion]
Posted by "Jefffrey (via GitHub)" <gi...@apache.org>.
Jefffrey commented on PR #20093:
URL: https://github.com/apache/datafusion/pull/20093#issuecomment-3832799154
Thanks @kumarUjjawal
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Refactor `iszero()` and `isnan()` to accept all numeric types [datafusion]
Posted by "Jefffrey (via GitHub)" <gi...@apache.org>.
Jefffrey commented on PR #20093:
URL: https://github.com/apache/datafusion/pull/20093#issuecomment-3828622953
For decimal is_zero we need to consider the scale
E.g. negative scale can never be zero
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Refactor `iszero()` and `isnan()` to accept all numeric types [datafusion]
Posted by "kumarUjjawal (via GitHub)" <gi...@apache.org>.
kumarUjjawal commented on code in PR #20093:
URL: https://github.com/apache/datafusion/pull/20093#discussion_r2749669561
##########
datafusion/functions/src/math/nans.rs:
##########
@@ -84,36 +88,123 @@ impl ScalarUDFImpl for IsNanFunc {
}
fn invoke_with_args(&self, args: ScalarFunctionArgs) -> Result<ColumnarValue> {
- // Handle NULL input
- if args.args[0].data_type().is_null() {
- return Ok(ColumnarValue::Scalar(ScalarValue::Boolean(None)));
- }
+ let [arg] = take_function_args(self.name(), args.args)?;
+
+ match arg {
+ ColumnarValue::Scalar(scalar) => {
+ if scalar.is_null() {
+ return Ok(ColumnarValue::Scalar(ScalarValue::Boolean(None)));
+ }
+
+ let result = match scalar {
+ ScalarValue::Float64(Some(v)) => Some(v.is_nan()),
+ ScalarValue::Float32(Some(v)) => Some(v.is_nan()),
+ ScalarValue::Float16(Some(v)) => Some(v.is_nan()),
- let args = ColumnarValue::values_to_arrays(&args.args)?;
-
- let arr: ArrayRef = match args[0].data_type() {
- DataType::Float64 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float64Type>(),
- f64::is_nan,
- )) as ArrayRef,
-
- DataType::Float32 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float32Type>(),
- f32::is_nan,
- )) as ArrayRef,
-
- DataType::Float16 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float16Type>(),
- |x| x.is_nan(),
- )) as ArrayRef,
- other => {
- return exec_err!(
- "Unsupported data type {other:?} for function {}",
- self.name()
- );
+ // Non-float numeric inputs are never NaN
+ ScalarValue::Int8(_)
+ | ScalarValue::Int16(_)
+ | ScalarValue::Int32(_)
+ | ScalarValue::Int64(_)
+ | ScalarValue::UInt8(_)
+ | ScalarValue::UInt16(_)
+ | ScalarValue::UInt32(_)
+ | ScalarValue::UInt64(_)
+ | ScalarValue::Decimal32(_, _, _)
+ | ScalarValue::Decimal64(_, _, _)
+ | ScalarValue::Decimal128(_, _, _)
+ | ScalarValue::Decimal256(_, _, _) => Some(false),
+
+ other => {
+ return exec_err!(
+ "Unsupported data type {other:?} for function {}",
+ self.name()
+ );
+ }
+ };
+
+ Ok(ColumnarValue::Scalar(ScalarValue::Boolean(result)))
}
- };
- Ok(ColumnarValue::Array(arr))
+ ColumnarValue::Array(array) => {
+ // NOTE: BooleanArray::from_unary preserves nulls.
+ let arr: ArrayRef = match array.data_type() {
+ Null => Arc::new(BooleanArray::new_null(array.len())) as ArrayRef,
+
+ Float64 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float64Type>(),
+ f64::is_nan,
+ )) as ArrayRef,
+ Float32 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float32Type>(),
+ f32::is_nan,
+ )) as ArrayRef,
+ Float16 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float16Type>(),
+ |x| x.is_nan(),
+ )) as ArrayRef,
+
+ // Non-float numeric arrays are never NaN
+ Decimal32(_, _) => Arc::new(BooleanArray::from_unary(
Review Comment:
Scalar NULL would get broadcast to all rows, so for a Decimal* array it would incorrectly return NULL even for non-null inputs. For decimal arrays we need false for non-null rows and NULL for null rows, so returning an array which preserves nulls) is required; scalar NULL is only correct for DataType::Null inputs ?
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Refactor `iszero()` and `isnan()` to accept all numeric types [datafusion]
Posted by "Jefffrey (via GitHub)" <gi...@apache.org>.
Jefffrey commented on code in PR #20093:
URL: https://github.com/apache/datafusion/pull/20093#discussion_r2749641023
##########
datafusion/functions/src/math/nans.rs:
##########
@@ -84,36 +88,123 @@ impl ScalarUDFImpl for IsNanFunc {
}
fn invoke_with_args(&self, args: ScalarFunctionArgs) -> Result<ColumnarValue> {
- // Handle NULL input
- if args.args[0].data_type().is_null() {
- return Ok(ColumnarValue::Scalar(ScalarValue::Boolean(None)));
- }
+ let [arg] = take_function_args(self.name(), args.args)?;
+
+ match arg {
+ ColumnarValue::Scalar(scalar) => {
+ if scalar.is_null() {
+ return Ok(ColumnarValue::Scalar(ScalarValue::Boolean(None)));
+ }
+
+ let result = match scalar {
+ ScalarValue::Float64(Some(v)) => Some(v.is_nan()),
+ ScalarValue::Float32(Some(v)) => Some(v.is_nan()),
+ ScalarValue::Float16(Some(v)) => Some(v.is_nan()),
- let args = ColumnarValue::values_to_arrays(&args.args)?;
-
- let arr: ArrayRef = match args[0].data_type() {
- DataType::Float64 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float64Type>(),
- f64::is_nan,
- )) as ArrayRef,
-
- DataType::Float32 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float32Type>(),
- f32::is_nan,
- )) as ArrayRef,
-
- DataType::Float16 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float16Type>(),
- |x| x.is_nan(),
- )) as ArrayRef,
- other => {
- return exec_err!(
- "Unsupported data type {other:?} for function {}",
- self.name()
- );
+ // Non-float numeric inputs are never NaN
+ ScalarValue::Int8(_)
+ | ScalarValue::Int16(_)
+ | ScalarValue::Int32(_)
+ | ScalarValue::Int64(_)
+ | ScalarValue::UInt8(_)
+ | ScalarValue::UInt16(_)
+ | ScalarValue::UInt32(_)
+ | ScalarValue::UInt64(_)
+ | ScalarValue::Decimal32(_, _, _)
+ | ScalarValue::Decimal64(_, _, _)
+ | ScalarValue::Decimal128(_, _, _)
+ | ScalarValue::Decimal256(_, _, _) => Some(false),
+
+ other => {
+ return exec_err!(
+ "Unsupported data type {other:?} for function {}",
+ self.name()
+ );
+ }
+ };
+
+ Ok(ColumnarValue::Scalar(ScalarValue::Boolean(result)))
}
- };
- Ok(ColumnarValue::Array(arr))
+ ColumnarValue::Array(array) => {
+ // NOTE: BooleanArray::from_unary preserves nulls.
+ let arr: ArrayRef = match array.data_type() {
+ Null => Arc::new(BooleanArray::new_null(array.len())) as ArrayRef,
+
+ Float64 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float64Type>(),
+ f64::is_nan,
+ )) as ArrayRef,
+ Float32 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float32Type>(),
+ f32::is_nan,
+ )) as ArrayRef,
+ Float16 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float16Type>(),
+ |x| x.is_nan(),
+ )) as ArrayRef,
+
+ // Non-float numeric arrays are never NaN
+ Decimal32(_, _) => Arc::new(BooleanArray::from_unary(
Review Comment:
For these array cases we can still return scalar null
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Refactor `iszero()` and `isnan()` to accept all numeric types [datafusion]
Posted by "Jefffrey (via GitHub)" <gi...@apache.org>.
Jefffrey commented on code in PR #20093:
URL: https://github.com/apache/datafusion/pull/20093#discussion_r2749670562
##########
datafusion/functions/src/math/nans.rs:
##########
@@ -84,36 +88,123 @@ impl ScalarUDFImpl for IsNanFunc {
}
fn invoke_with_args(&self, args: ScalarFunctionArgs) -> Result<ColumnarValue> {
- // Handle NULL input
- if args.args[0].data_type().is_null() {
- return Ok(ColumnarValue::Scalar(ScalarValue::Boolean(None)));
- }
+ let [arg] = take_function_args(self.name(), args.args)?;
+
+ match arg {
+ ColumnarValue::Scalar(scalar) => {
+ if scalar.is_null() {
+ return Ok(ColumnarValue::Scalar(ScalarValue::Boolean(None)));
+ }
+
+ let result = match scalar {
+ ScalarValue::Float64(Some(v)) => Some(v.is_nan()),
+ ScalarValue::Float32(Some(v)) => Some(v.is_nan()),
+ ScalarValue::Float16(Some(v)) => Some(v.is_nan()),
- let args = ColumnarValue::values_to_arrays(&args.args)?;
-
- let arr: ArrayRef = match args[0].data_type() {
- DataType::Float64 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float64Type>(),
- f64::is_nan,
- )) as ArrayRef,
-
- DataType::Float32 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float32Type>(),
- f32::is_nan,
- )) as ArrayRef,
-
- DataType::Float16 => Arc::new(BooleanArray::from_unary(
- args[0].as_primitive::<Float16Type>(),
- |x| x.is_nan(),
- )) as ArrayRef,
- other => {
- return exec_err!(
- "Unsupported data type {other:?} for function {}",
- self.name()
- );
+ // Non-float numeric inputs are never NaN
+ ScalarValue::Int8(_)
+ | ScalarValue::Int16(_)
+ | ScalarValue::Int32(_)
+ | ScalarValue::Int64(_)
+ | ScalarValue::UInt8(_)
+ | ScalarValue::UInt16(_)
+ | ScalarValue::UInt32(_)
+ | ScalarValue::UInt64(_)
+ | ScalarValue::Decimal32(_, _, _)
+ | ScalarValue::Decimal64(_, _, _)
+ | ScalarValue::Decimal128(_, _, _)
+ | ScalarValue::Decimal256(_, _, _) => Some(false),
+
+ other => {
+ return exec_err!(
+ "Unsupported data type {other:?} for function {}",
+ self.name()
+ );
+ }
+ };
+
+ Ok(ColumnarValue::Scalar(ScalarValue::Boolean(result)))
}
- };
- Ok(ColumnarValue::Array(arr))
+ ColumnarValue::Array(array) => {
+ // NOTE: BooleanArray::from_unary preserves nulls.
+ let arr: ArrayRef = match array.data_type() {
+ Null => Arc::new(BooleanArray::new_null(array.len())) as ArrayRef,
+
+ Float64 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float64Type>(),
+ f64::is_nan,
+ )) as ArrayRef,
+ Float32 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float32Type>(),
+ f32::is_nan,
+ )) as ArrayRef,
+ Float16 => Arc::new(BooleanArray::from_unary(
+ array.as_primitive::<Float16Type>(),
+ |x| x.is_nan(),
+ )) as ArrayRef,
+
+ // Non-float numeric arrays are never NaN
+ Decimal32(_, _) => Arc::new(BooleanArray::from_unary(
Review Comment:
Ah you're right, my mistake
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] Refactor `iszero()` and `isnan()` to accept all numeric types [datafusion]
Posted by "Jefffrey (via GitHub)" <gi...@apache.org>.
Jefffrey merged PR #20093:
URL: https://github.com/apache/datafusion/pull/20093
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org