comphead commented on code in PR #74:
URL: https://github.com/apache/datafusion-spark/pull/74#discussion_r4566785952
##########
crates/spark/src/function/datetime/date_from_unix_date.rs:
##########
@@ -0,0 +1,107 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: See the NOTICE file distributed with this work for
additional information regarding copyright ownership
+// SPDX-FileContributor: Licensed to the Apache Software Foundation (ASF)
under one or more contributor license agreements
+
+use arrow::compute::cast;
+use arrow::datatypes::DataType;
+use datafusion_common::types::{NativeType, logical_int32};
+use datafusion_common::utils::take_function_args;
+use datafusion_common::{Result, ScalarValue, internal_err};
+use datafusion_expr::{
+ Coercion, ColumnarValue, ScalarFunctionArgs, ScalarUDFImpl, Signature,
+ TypeSignatureClass, Volatility,
+};
+
+/// Spark-compatible `date_from_unix_date(days)`.
+///
<https://spark.apache.org/docs/latest/api/sql/index.html#date_from_unix_date>
+///
+/// Creates a date from the number of days since 1970-01-01. The argument is an
+/// INT (other integer types are cast to it), and NULL gives NULL. Any INT is a
+/// valid date, so there is no error case. This is the inverse of `unix_date`.
+#[derive(Debug, PartialEq, Eq, Hash)]
+pub struct SparkDateFromUnixDate {
+ signature: Signature,
+}
+
+impl Default for SparkDateFromUnixDate {
+ fn default() -> Self {
+ Self::new()
+ }
+}
+
+impl SparkDateFromUnixDate {
+ pub fn new() -> Self {
+ let int32 = Coercion::new_implicit(
+ TypeSignatureClass::Native(logical_int32()),
+ vec![TypeSignatureClass::Integer],
+ NativeType::Int32,
+ );
+ Self {
+ signature: Signature::coercible(vec![int32],
Volatility::Immutable),
+ }
+ }
+}
+
+impl ScalarUDFImpl for SparkDateFromUnixDate {
+ fn name(&self) -> &str {
+ "date_from_unix_date"
+ }
+
+ fn signature(&self) -> &Signature {
+ &self.signature
+ }
+
+ fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
+ Ok(DataType::Date32)
+ }
+
+ fn invoke_with_args(&self, args: ScalarFunctionArgs) ->
Result<ColumnarValue> {
+ let [days] = take_function_args(self.name(), args.args)?;
+ match days {
+ // Date32 is Int32 days since the epoch, so this cast reuses the
buffer.
+ ColumnarValue::Array(array) => {
+ Ok(ColumnarValue::Array(cast(&array, &DataType::Date32)?))
+ }
+ ColumnarValue::Scalar(ScalarValue::Int32(days)) => {
+ Ok(ColumnarValue::Scalar(ScalarValue::Date32(days)))
+ }
+ ColumnarValue::Scalar(other) => {
+ internal_err!(
+ "date_from_unix_date expects INT, got {}",
+ other.data_type()
+ )
+ }
+ }
+ }
+}
+
+#[cfg(test)]
Review Comment:
`days_become_dates` checks 0, 1000, NULL and -1, which
`date_from_unix_date.slt` already covers, including the column path. CI runs
`cargo test -p datafusion-spark-sqllogictest` but not `cargo test -p
datafusion-spark`, so Rust unit tests do not run in CI. Could we drop this
module?
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]