comphead commented on code in PR #8528:
URL: https://github.com/apache/arrow-datafusion/pull/8528#discussion_r1425646646
##########
datafusion/physical-expr/src/datetime_expressions.rs:
##########
@@ -647,89 +644,94 @@ fn date_bin_impl(
return exec_err!("DATE_BIN stride must be non-zero");
}
- let f_nanos = |x: Option<i64>| x.map(|x| stride_fn(stride, x, origin));
- let f_micros = |x: Option<i64>| {
- let scale = 1_000;
- x.map(|x| stride_fn(stride, x * scale, origin) / scale)
- };
- let f_millis = |x: Option<i64>| {
- let scale = 1_000_000;
- x.map(|x| stride_fn(stride, x * scale, origin) / scale)
- };
- let f_secs = |x: Option<i64>| {
- let scale = 1_000_000_000;
- x.map(|x| stride_fn(stride, x * scale, origin) / scale)
- };
+ fn stride_map_fn<T: ArrowTimestampType>(
+ origin: i64,
+ stride: i64,
+ stride_fn: fn(i64, i64, i64) -> i64,
+ ) -> impl Fn(Option<i64>) -> Option<i64> {
+ let scale = match T::UNIT {
+ TimeUnit::Nanosecond => 1,
+ TimeUnit::Microsecond => NANOSECONDS / 1_000_000,
+ TimeUnit::Millisecond => NANOSECONDS / 1_000,
+ TimeUnit::Second => NANOSECONDS,
+ };
+ move |x: Option<i64>| x.map(|x| stride_fn(stride, x * scale, origin) /
scale)
+ }
Ok(match array {
ColumnarValue::Scalar(ScalarValue::TimestampNanosecond(v, tz_opt)) => {
- ColumnarValue::Scalar(ScalarValue::TimestampNanosecond(
- f_nanos(*v),
- tz_opt.clone(),
- ))
+ let f = stride_map_fn::<TimestampNanosecondType>(origin, stride,
stride_fn);
+ ColumnarValue::Scalar(ScalarValue::TimestampNanosecond(f(*v),
tz_opt.clone()))
}
ColumnarValue::Scalar(ScalarValue::TimestampMicrosecond(v, tz_opt)) =>
{
+ let f = stride_map_fn::<TimestampMicrosecondType>(origin, stride,
stride_fn);
ColumnarValue::Scalar(ScalarValue::TimestampMicrosecond(
- f_micros(*v),
+ f(*v),
tz_opt.clone(),
))
}
ColumnarValue::Scalar(ScalarValue::TimestampMillisecond(v, tz_opt)) =>
{
+ let f = stride_map_fn::<TimestampMillisecondType>(origin, stride,
stride_fn);
ColumnarValue::Scalar(ScalarValue::TimestampMillisecond(
- f_millis(*v),
+ f(*v),
tz_opt.clone(),
))
}
ColumnarValue::Scalar(ScalarValue::TimestampSecond(v, tz_opt)) => {
- ColumnarValue::Scalar(ScalarValue::TimestampSecond(
- f_secs(*v),
- tz_opt.clone(),
- ))
+ let f = stride_map_fn::<TimestampSecondType>(origin, stride,
stride_fn);
+ ColumnarValue::Scalar(ScalarValue::TimestampSecond(f(*v),
tz_opt.clone()))
}
- ColumnarValue::Array(array) => match array.data_type() {
- DataType::Timestamp(TimeUnit::Nanosecond, tz_opt) => {
- let array = as_timestamp_nanosecond_array(array)?
- .iter()
- .map(f_nanos)
- .collect::<TimestampNanosecondArray>()
- .with_timezone_opt(tz_opt.clone());
-
- ColumnarValue::Array(Arc::new(array))
- }
- DataType::Timestamp(TimeUnit::Microsecond, tz_opt) => {
- let array = as_timestamp_microsecond_array(array)?
- .iter()
- .map(f_micros)
- .collect::<TimestampMicrosecondArray>()
- .with_timezone_opt(tz_opt.clone());
- ColumnarValue::Array(Arc::new(array))
- }
- DataType::Timestamp(TimeUnit::Millisecond, tz_opt) => {
- let array = as_timestamp_millisecond_array(array)?
- .iter()
- .map(f_millis)
- .collect::<TimestampMillisecondArray>()
- .with_timezone_opt(tz_opt.clone());
-
- ColumnarValue::Array(Arc::new(array))
- }
- DataType::Timestamp(TimeUnit::Second, tz_opt) => {
- let array = as_timestamp_second_array(array)?
+ ColumnarValue::Array(array) => {
+ fn process_array<T>(
Review Comment:
can we name a function more precisely, `process_array` is too generic and
requires the reader go to the functions internal code to figure what exactly it
does
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
