Skip to content

Centralize DATE_BIN source scaling and binning through shared helper #22823

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
kosiew wants to merge 15 commits into
base: main
Choose a base branch
from
+34 −25
Open

Centralize DATE_BIN source scaling and binning through shared helper #22823

Changes from all commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
9e20bb1
feat(datetime): add overflow reproducer for Time64Microsecond and Tim…
kosiew Jun 8, 2026
3e365dc
feat(date_bin): improve datetime functions with value_to_nanos and op…
kosiew Jun 8, 2026
60caf63
feat: update TIME scaling to use value_to_nanos in date_bin.rs
kosiew Jun 8, 2026
25394b7
feat(datetime): rename overflow helper and update error handling in d…
kosiew Jun 8, 2026
8feac1b
fix: update expected timestamp overflow error in date_bin_errors.slt …
kosiew Jun 8, 2026
37f6b0d
feat(datetime): restore timestamp overflow text and improve error han…
kosiew Jun 8, 2026
2921e86
fix(datetime): clarify error messages for DATE_BIN source timestamp o…
kosiew Jun 8, 2026
a991f68
feat(date_bin): refactor date_bin.rs and date_bin_errors.slt
kosiew Jun 8, 2026
e99bc47
Merge branch 'main' into scaling-overflow-02-22685
kosiew Jun 17, 2026
a0d6684
fix(datetime): replace integer multiplication with checked_scale_to_n…
kosiew Jun 17, 2026
2d6931e
feat(datetime): rename helper argument and enhance scaling and binnin…
kosiew Jun 17, 2026
7d596e1
fix(datetime): revert Time64(Nanosecond) array branch to preserve err…
kosiew Jun 17, 2026
c2fd024
feat: rename checked_scale_and_bin_to_nanos to checked_scale_and_bin_…
kosiew Jun 17, 2026
3ddd544
feat(datetime): format code in date_bin.rs for improved readability
kosiew Jun 17, 2026
bb4807e
feat: refactor scaling functions and improve error handling
kosiew Jun 20, 2026
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
59 changes: 34 additions & 25 deletions datafusion/functions/src/datetime/date_bin.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -437,10 +437,12 @@ fn timestamp_scale<T: ArrowTimestampType>() -> i64 {
}

// Scale to nanoseconds and report overflow as a normal error.
fn checked_scale_to_nanos(x: i64, scale: i64) -> Result<i64> {
match x.checked_mul(scale) {
fn checked_scale_to_nanos(value: i64, scale: i64) -> Result<i64> {
match value.checked_mul(scale) {
Some(scaled) => Ok(scaled),
None => exec_err!("date_bin timestamp value {x} * scale {scale} overflows i64"),
None => {
exec_err!("date_bin timestamp value {value} * scale {scale} overflows i64")
}
}
}

Expand Down Expand Up @@ -570,7 +572,7 @@ fn date_bin_impl(
) -> Option<i64> {
let scale = timestamp_scale::<T>();
value
.and_then(|val| val.checked_mul(scale))
.and_then(|value| checked_scale_to_nanos(value, scale).ok())
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| binned / scale)
}
Expand Down Expand Up @@ -613,7 +615,9 @@ fn date_bin_impl(
return exec_err!("DATE_BIN with Time32 source requires Time32 origin");
}
let result = v
.and_then(|x| (x as i64).checked_mul(NANOS_PER_MILLI))
.and_then(|value| {
checked_scale_to_nanos(value as i64, NANOS_PER_MILLI).ok()
})
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| ((binned % NANOSECONDS_IN_DAY) / NANOS_PER_MILLI) as i32);
ColumnarValue::Scalar(ScalarValue::Time32Millisecond(result))
Expand All @@ -623,7 +627,9 @@ fn date_bin_impl(
return exec_err!("DATE_BIN with Time32 source requires Time32 origin");
}
let result = v
.and_then(|x| (x as i64).checked_mul(NANOS_PER_SEC))
.and_then(|value| {
checked_scale_to_nanos(value as i64, NANOS_PER_SEC).ok()
})
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| ((binned % NANOSECONDS_IN_DAY) / NANOS_PER_SEC) as i32);
ColumnarValue::Scalar(ScalarValue::Time32Second(result))
Expand All @@ -644,7 +650,7 @@ fn date_bin_impl(
return exec_err!("DATE_BIN with Time64 source requires Time64 origin");
}
let result = v
.and_then(|x| x.checked_mul(NANOS_PER_MICRO))
.and_then(|value| checked_scale_to_nanos(value, NANOS_PER_MICRO).ok())
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| (binned % NANOSECONDS_IN_DAY) / NANOS_PER_MICRO);
ColumnarValue::Scalar(ScalarValue::Time64Microsecond(result))
Expand All @@ -664,8 +670,9 @@ fn date_bin_impl(
let scale = timestamp_scale::<T>();

// Per-row errors become NULL, matching scalar behavior.
let result: PrimitiveArray<T> = array.unary_opt(|val| {
val.checked_mul(scale)
let result: PrimitiveArray<T> = array.unary_opt(|value| {
checked_scale_to_nanos(value, scale)
.ok()
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| binned / scale)
});
Expand Down Expand Up @@ -703,9 +710,9 @@ fn date_bin_impl(
}
let array = array.as_primitive::<Time32MillisecondType>();
let result: PrimitiveArray<Time32MillisecondType> =
array.unary_opt(|x| {
(x as i64)
.checked_mul(NANOS_PER_MILLI)
array.unary_opt(|value| {
checked_scale_to_nanos(value as i64, NANOS_PER_MILLI)
.ok()
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| {
((binned % NANOSECONDS_IN_DAY) / NANOS_PER_MILLI)
Expand All @@ -721,14 +728,15 @@ fn date_bin_impl(
);
}
let array = array.as_primitive::<Time32SecondType>();
let result: PrimitiveArray<Time32SecondType> = array.unary_opt(|x| {
(x as i64)
.checked_mul(NANOS_PER_SEC)
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| {
((binned % NANOSECONDS_IN_DAY) / NANOS_PER_SEC) as i32
})
});
let result: PrimitiveArray<Time32SecondType> =
array.unary_opt(|value| {
checked_scale_to_nanos(value as i64, NANOS_PER_SEC)
.ok()
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| {
((binned % NANOSECONDS_IN_DAY) / NANOS_PER_SEC) as i32
})
});
ColumnarValue::Array(Arc::new(result))
}
Time64(Microsecond) => {
Expand All @@ -739,8 +747,9 @@ fn date_bin_impl(
}
let array = array.as_primitive::<Time64MicrosecondType>();
let result: PrimitiveArray<Time64MicrosecondType> =
array.unary_opt(|x| {
x.checked_mul(NANOS_PER_MICRO)
array.unary_opt(|value| {
checked_scale_to_nanos(value, NANOS_PER_MICRO)
.ok()
.and_then(|scaled| stride_fn(stride, scaled, origin).ok())
.map(|binned| {
(binned % NANOSECONDS_IN_DAY) / NANOS_PER_MICRO
Expand All @@ -756,11 +765,11 @@ fn date_bin_impl(
}
let array = array.as_primitive::<Time64NanosecondType>();
let result: PrimitiveArray<Time64NanosecondType> =
array.unary_opt(|x| {
array.try_unary(|x| {
stride_fn(stride, x, origin)
.map(|binned_nanos| binned_nanos % (NANOSECONDS_IN_DAY))
.ok()
});
.map_err(|e| ArrowError::ComputeError(e.to_string()))
})?;
ColumnarValue::Array(Arc::new(result))
}
_ => {
Expand Down
Loading