From 2d2ff1be6416aef564cb67b6c6682eceb22b99df Mon Sep 17 00:00:00 2001 From: Tim Liu Date: Wed, 25 Oct 2023 16:38:39 +0800 Subject: [PATCH] Add pyarrow tests for Databricks runtime (#9534) To fix: https://github.com/NVIDIA/spark-rapids/issues/9533 Signed-off-by: Tim Liu --- jenkins/databricks/test.sh | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/jenkins/databricks/test.sh b/jenkins/databricks/test.sh index 6c96e45ff72..404dcd97578 100755 --- a/jenkins/databricks/test.sh +++ b/jenkins/databricks/test.sh @@ -58,6 +58,7 @@ IS_SPARK_321_OR_LATER=0 # - DEFAULT: all tests except cudf_udf tests # - DELTA_LAKE_ONLY: delta_lake tests only # - MULTITHREADED_SHUFFLE: shuffle tests only +# - PYARROW_ONLY: pyarrow tests only TEST_MODE=${TEST_MODE:-'DEFAULT'} # Classloader config is here to work around classloader issues with @@ -83,6 +84,11 @@ rapids_shuffle_smoke_test() { bash integration_tests/run_pyspark_from_build.sh -m shuffle_test --runtime_env="databricks" --test_type=$TEST_TYPE } +run_pyarrow_tests() { + SPARK_SUBMIT_FLAGS="$SPARK_CONF" \ + bash integration_tests/run_pyspark_from_build.sh -m pyarrow_test --pyarrow_test --runtime_env="databricks" --test_type=$TEST_TYPE +} + ## limit parallelism to avoid OOM kill export TEST_PARALLEL=${TEST_PARALLEL:-4} @@ -108,4 +114,8 @@ if [ "$(pwd)" == "$SOURCE_PATH" ]; then ## Mutithreaded Shuffle test rapids_shuffle_smoke_test fi + if [[ "$TEST_MODE" == "DEFAULT" || "$TEST_MODE" == "PYARROW_ONLY" ]]; then + # Pyarrow tests + run_pyarrow_tests + fi fi