You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "allisonwang-db (via GitHub)" <gi...@apache.org> on 2023/07/03 21:05:09 UTC
[GitHub] [spark] allisonwang-db commented on a diff in pull request #41606: [WIP] [SPARK-44061] Add assertDFEqual util function
allisonwang-db commented on code in PR #41606:
URL: https://github.com/apache/spark/pull/41606#discussion_r1251258342
##########
python/pyspark/testing/utils.py:
##########
@@ -209,3 +216,50 @@ def check_error(
self.assertEqual(
expected, actual, f"Expected message parameters was '{expected}', got '{actual}'"
)
+
+
+def assertSparkSchemaEquality(
+ s1: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+ s2: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+):
+ if s1 != s2:
+ msg = "Schemas are different"
+ raise AssertionError(msg)
+
+
+def assertSparkDFEqual(left: PySparkDataFrame, right: PySparkDataFrame):
+ def assert_rows_equality(rows1, rows2):
+ if rows1 != rows2:
Review Comment:
Let's add some tests for
1. None/Null
2. Complex types (array, map, struct)
3. Floating points
##########
python/pyspark/testing/utils.py:
##########
@@ -209,3 +216,50 @@ def check_error(
self.assertEqual(
expected, actual, f"Expected message parameters was '{expected}', got '{actual}'"
)
+
+
+def assertSparkSchemaEquality(
+ s1: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+ s2: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+):
+ if s1 != s2:
+ msg = "Schemas are different"
+ raise AssertionError(msg)
+
+
+def assertSparkDFEqual(left: PySparkDataFrame, right: PySparkDataFrame):
+ def assert_rows_equality(rows1, rows2):
+ if rows1 != rows2:
+ raise PySparkAssertionError(
+ error_class="DIFFERENT_DATAFRAME",
+ message_parameters={},
+ )
+
+ left = left.sort(left.columns)
Review Comment:
We don't always want the rows to be sorted, e.g. when a query has a "ORDER BY" clause. We can check the logical plan of the dataframe to see if there are any `Sort`.
```
val isSorted = df.logicalPlan.collect { case s: logical.Sort => s }.nonEmpty
```
I am not sure if this works for Spark Connect tests.
##########
python/pyspark/testing/utils.py:
##########
@@ -209,3 +216,50 @@ def check_error(
self.assertEqual(
expected, actual, f"Expected message parameters was '{expected}', got '{actual}'"
)
+
+
+def assertSparkSchemaEquality(
Review Comment:
assertSchemaEqual
##########
python/pyspark/testing/utils.py:
##########
@@ -209,3 +216,50 @@ def check_error(
self.assertEqual(
expected, actual, f"Expected message parameters was '{expected}', got '{actual}'"
)
+
+
+def assertSparkSchemaEquality(
+ s1: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+ s2: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+):
+ if s1 != s2:
+ msg = "Schemas are different"
+ raise AssertionError(msg)
+
+
+def assertSparkDFEqual(left: PySparkDataFrame, right: PySparkDataFrame):
Review Comment:
Maybe we can have this:
assertDataFrameEqual(df: DataFrame, expected: Union[DataFrame, List[Row]])
We can use the `checkAnswer` from the Scala side for reference.
##########
python/pyspark/testing/utils.py:
##########
@@ -209,3 +216,50 @@ def check_error(
self.assertEqual(
expected, actual, f"Expected message parameters was '{expected}', got '{actual}'"
)
+
+
+def assertSparkSchemaEquality(
+ s1: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+ s2: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+):
+ if s1 != s2:
+ msg = "Schemas are different"
+ raise AssertionError(msg)
+
+
+def assertSparkDFEqual(left: PySparkDataFrame, right: PySparkDataFrame):
+ def assert_rows_equality(rows1, rows2):
+ if rows1 != rows2:
+ raise PySparkAssertionError(
+ error_class="DIFFERENT_DATAFRAME",
+ message_parameters={},
+ )
+
+ left = left.sort(left.columns)
Review Comment:
If checking the logical plan is not doable, we can probably have a param `ignore_row_order=True` for users to specify whether they want the rows sorted.
##########
python/pyspark/testing/utils.py:
##########
@@ -209,3 +216,50 @@ def check_error(
self.assertEqual(
expected, actual, f"Expected message parameters was '{expected}', got '{actual}'"
)
+
+
+def assertSparkSchemaEquality(
+ s1: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+ s2: Optional[Union[AtomicType, StructType, str, List[str], Tuple[str, ...]]],
+):
+ if s1 != s2:
+ msg = "Schemas are different"
+ raise AssertionError(msg)
+
+
+def assertSparkDFEqual(left: PySparkDataFrame, right: PySparkDataFrame):
+ def assert_rows_equality(rows1, rows2):
+ if rows1 != rows2:
+ raise PySparkAssertionError(
Review Comment:
Do we need to use error classes here?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org