lidavidm commented on a change in pull request #11313:
URL: https://github.com/apache/arrow/pull/11313#discussion_r742834941
##########
File path: cpp/src/arrow/compute/kernels/aggregate_test.cc
##########
@@ -3541,6 +3649,24 @@ TEST(TestQuantileKernel, AllNullsOrNaNs) {
}
}
+TEST(TestQuantileKernel, Decimal) {
+ auto check = [](const std::shared_ptr<Array>& input, QuantileOptions options,
+ const std::shared_ptr<Array>& expected) {
+ ASSERT_OK_AND_ASSIGN(Datum out, Quantile(input, options));
+ auto out_array = out.make_array();
+ ValidateOutput(*out_array);
+ AssertArraysEqual(*expected, *out_array, /*verbose=*/true);
+ };
+ for (const auto& ty : {decimal128(3, 2), decimal256(3, 2)}) {
+ check(ArrayFromJSON(ty, R"(["1.00", "5.00", null])"),
+ QuantileOptions(0.5, QuantileOptions::LINEAR),
+ ArrayFromJSON(float64(), R"([3.00])"));
Review comment:
We'd then have to convert the quantile `q` to decimal (and accept
truncation or promote the precision in the output type). IMO, it's more
consistent for interpolated quantiles to always be a float.
--
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]
