diff --git a/integration_tests/run_pyspark_from_build.sh b/integration_tests/run_pyspark_from_build.sh index 90d05e55ae3..017794b4167 100755 --- a/integration_tests/run_pyspark_from_build.sh +++ b/integration_tests/run_pyspark_from_build.sh @@ -527,11 +527,12 @@ else exit 1 fi - # Create a venv and install only pyspark[connect] to ensure a pure Python client + # Create a venv and install only pyspark-client to ensure a pure Python client + # See: https://spark.apache.org/docs/latest/api/python/getting_started/install.html#python-spark-connect-client CONNECT_CLIENT_VENV="${RUN_DIR}/connect_client_venv" python -m venv "$CONNECT_CLIENT_VENV" "$CONNECT_CLIENT_VENV/bin/python" -m pip install --upgrade pip >/dev/null - "$CONNECT_CLIENT_VENV/bin/python" -m pip install --no-cache-dir "pyspark[connect]==${VERSION_STRING}" >/dev/null + "$CONNECT_CLIENT_VENV/bin/python" -m pip install --no-cache-dir "pyspark-client==${VERSION_STRING}" > /dev/null # Run a simple query using the Connect client and assert expected result and GPU operator in the plan output=$(CONNECT_URL="$CONNECT_SERVER_URL" \