[SPARK-3786] [PySpark] speedup tests

This patch try to speed up tests of PySpark, re-use the SparkContext in tests.py and mllib/tests.py to reduce the overhead of create SparkContext, remove some test cases, which did not make sense. It also improve the performance of some cases, such as MergerTests and SortTests.

before this patch:

real	21m27.320s
user	4m42.967s
sys	0m17.343s

after this patch:

real	9m47.541s
user	2m12.947s
sys	0m14.543s

It almost cut the time by half.

Author: Davies Liu <davies.liu@gmail.com>

Closes #2646 from davies/tests and squashes the following commits:

c54de60 [Davies Liu] revert change about memory limit
6a2a4b0 [Davies Liu] refactor of tests, speedup 100%

1 files changed, 37 insertions, 37 deletions

diff --git a/python/run-tests b/python/run-tests
index a7ec270c7d..c713861eb7 100755
--- a/python/run-tests
+++ b/python/run-tests
@@ -34,7 +34,7 @@ rm -rf metastore warehouse
 function run_test() {
     echo "Running test: $1"
 
-    SPARK_TESTING=1 "$FWDIR"/bin/pyspark $1 2>&1 | tee -a unit-tests.log
+    SPARK_TESTING=1 time "$FWDIR"/bin/pyspark $1 2>&1 | tee -a unit-tests.log
 
     FAILED=$((PIPESTATUS[0]||$FAILED))
 
@@ -48,6 +48,37 @@ function run_test() {
     fi
 }
 
+function run_core_tests() {
+    echo "Run core tests ..."
+    run_test "pyspark/rdd.py"
+    run_test "pyspark/context.py"
+    run_test "pyspark/conf.py"
+    PYSPARK_DOC_TEST=1 run_test "pyspark/broadcast.py"
+    PYSPARK_DOC_TEST=1 run_test "pyspark/accumulators.py"
+    PYSPARK_DOC_TEST=1 run_test "pyspark/serializers.py"
+    run_test "pyspark/shuffle.py"
+    run_test "pyspark/tests.py"
+}
+
+function run_sql_tests() {
+    echo "Run sql tests ..."
+    run_test "pyspark/sql.py"
+}
+
+function run_mllib_tests() {
+    echo "Run mllib tests ..."
+    run_test "pyspark/mllib/classification.py"
+    run_test "pyspark/mllib/clustering.py"
+    run_test "pyspark/mllib/linalg.py"
+    run_test "pyspark/mllib/random.py"
+    run_test "pyspark/mllib/recommendation.py"
+    run_test "pyspark/mllib/regression.py"
+    run_test "pyspark/mllib/stat.py"
+    run_test "pyspark/mllib/tree.py"
+    run_test "pyspark/mllib/util.py"
+    run_test "pyspark/mllib/tests.py"
+}
+
 echo "Running PySpark tests. Output is in python/unit-tests.log."
 
 export PYSPARK_PYTHON="python"
@@ -60,29 +91,9 @@ fi
 echo "Testing with Python version:"
 $PYSPARK_PYTHON --version
 
-run_test "pyspark/rdd.py"
-run_test "pyspark/context.py"
-run_test "pyspark/conf.py"
-run_test "pyspark/sql.py"
-# These tests are included in the module-level docs, and so must
-# be handled on a higher level rather than within the python file.
-export PYSPARK_DOC_TEST=1
-run_test "pyspark/broadcast.py"
-run_test "pyspark/accumulators.py"
-run_test "pyspark/serializers.py"
-unset PYSPARK_DOC_TEST
-run_test "pyspark/shuffle.py"
-run_test "pyspark/tests.py"
-run_test "pyspark/mllib/classification.py"
-run_test "pyspark/mllib/clustering.py"
-run_test "pyspark/mllib/linalg.py"
-run_test "pyspark/mllib/random.py"
-run_test "pyspark/mllib/recommendation.py"
-run_test "pyspark/mllib/regression.py"
-run_test "pyspark/mllib/stat.py"
-run_test "pyspark/mllib/tests.py"
-run_test "pyspark/mllib/tree.py"
-run_test "pyspark/mllib/util.py"
+run_core_tests
+run_sql_tests
+run_mllib_tests
 
 # Try to test with PyPy
 if [ $(which pypy) ]; then
@@ -90,19 +101,8 @@ if [ $(which pypy) ]; then
     echo "Testing with PyPy version:"
     $PYSPARK_PYTHON --version
 
-    run_test "pyspark/rdd.py"
-    run_test "pyspark/context.py"
-    run_test "pyspark/conf.py"
-    run_test "pyspark/sql.py"
-    # These tests are included in the module-level docs, and so must
-    # be handled on a higher level rather than within the python file.
-    export PYSPARK_DOC_TEST=1
-    run_test "pyspark/broadcast.py"
-    run_test "pyspark/accumulators.py"
-    run_test "pyspark/serializers.py"
-    unset PYSPARK_DOC_TEST
-    run_test "pyspark/shuffle.py"
-    run_test "pyspark/tests.py"
+    run_core_tests
+    run_sql_tests
 fi
 
 if [[ $FAILED == 0 ]]; then