From 91d93dbb1d2c145f8a9057975beca68f3141f4d9 Mon Sep 17 00:00:00 2001 From: Andy Grove Date: Fri, 12 Jun 2026 07:33:14 -0600 Subject: [PATCH] feat: opt sort_array into codegen dispatch under strict floating-point mode CometSortArray reports Incompatible only when spark.comet.exec.strictFloatingPoint is enabled and the array contains floats. Mixing in CodegenDispatchFallback routes that case through the JVM codegen dispatcher (Spark's own doGenCode) so it stays native and matches Spark instead of falling back. SortArray has a real doGenCode (not CodegenFallback), so the dispatcher accepts it. Part of #4596. --- .../scala/org/apache/comet/serde/arrays.scala | 2 +- .../array/sort_array_strict_fp.sql | 57 +++++++++++++++++++ 2 files changed, 58 insertions(+), 1 deletion(-) create mode 100644 spark/src/test/resources/sql-tests/expressions/array/sort_array_strict_fp.sql diff --git a/spark/src/main/scala/org/apache/comet/serde/arrays.scala b/spark/src/main/scala/org/apache/comet/serde/arrays.scala index 690bc376c7..f75acedc20 100644 --- a/spark/src/main/scala/org/apache/comet/serde/arrays.scala +++ b/spark/src/main/scala/org/apache/comet/serde/arrays.scala @@ -119,7 +119,7 @@ object CometArrayContains extends CometExpressionSerde[ArrayContains] { } } -object CometSortArray extends CometExpressionSerde[SortArray] { +object CometSortArray extends CometExpressionSerde[SortArray] with CodegenDispatchFallback { override def getIncompatibleReasons(): Seq[String] = Seq( "When `" + CometConf.COMET_EXEC_STRICT_FLOATING_POINT.key + "=true`, sorting on" + diff --git a/spark/src/test/resources/sql-tests/expressions/array/sort_array_strict_fp.sql b/spark/src/test/resources/sql-tests/expressions/array/sort_array_strict_fp.sql new file mode 100644 index 0000000000..29e27a7422 --- /dev/null +++ b/spark/src/test/resources/sql-tests/expressions/array/sort_array_strict_fp.sql @@ -0,0 +1,57 @@ +-- Licensed to the Apache Software Foundation (ASF) under one +-- or more contributor license agreements. See the NOTICE file +-- distributed with this work for additional information +-- regarding copyright ownership. The ASF licenses this file +-- to you under the Apache License, Version 2.0 (the +-- "License"); you may not use this file except in compliance +-- with the License. You may obtain a copy of the License at +-- +-- http://www.apache.org/licenses/LICENSE-2.0 +-- +-- Unless required by applicable law or agreed to in writing, +-- software distributed under the License is distributed on an +-- "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +-- KIND, either express or implied. See the License for the +-- specific language governing permissions and limitations +-- under the License. + +-- With strict floating-point mode on, Comet's native float sort is Incompatible with Spark +-- (NaN / signed-zero ordering). CometSortArray mixes in CodegenDispatchFallback, so it routes +-- through the JVM codegen dispatcher and matches Spark exactly instead of falling back. + +-- Config: spark.comet.exec.strictFloatingPoint=true + +statement +CREATE TABLE test_sort_array_strict_double(arr array) USING parquet + +statement +INSERT INTO test_sort_array_strict_double VALUES + (array(CAST('Infinity' AS DOUBLE), CAST('-Infinity' AS DOUBLE), CAST('NaN' AS DOUBLE), 3.0, 1.0, NULL, -0.0, 0.0)), + (array(CAST('NaN' AS DOUBLE), CAST('NaN' AS DOUBLE), -5.0, 2.0)), + (array()), + (NULL) + +query +SELECT sort_array(arr) FROM test_sort_array_strict_double + +query +SELECT sort_array(arr, false) FROM test_sort_array_strict_double + +statement +CREATE TABLE test_sort_array_strict_float(arr array) USING parquet + +statement +INSERT INTO test_sort_array_strict_float VALUES + (array(CAST('Infinity' AS FLOAT), CAST('-Infinity' AS FLOAT), CAST('NaN' AS FLOAT), CAST(3.0 AS FLOAT), CAST(NULL AS FLOAT), CAST(-0.0 AS FLOAT), CAST(0.0 AS FLOAT))), + (array()), + (NULL) + +query +SELECT sort_array(arr) FROM test_sort_array_strict_float + +query +SELECT sort_array(arr, false) FROM test_sort_array_strict_float + +-- nested float arrays also dispatch +query +SELECT sort_array(array(CAST('NaN' AS DOUBLE), -0.0, 0.0, 1.0, CAST('-Infinity' AS DOUBLE)))