[SPARK-41899] DataFrame.createDataFrame converting int to bigint

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.0
• Fix Version/s: 3.4.0
• Component/s: Connect
• Assignee: Ruifeng Zheng
• Reporter: Sandeep Singh
• Created: 05/Jan/23 17:34
• Updated: 08/Jan/23 05:03
• Resolved: 08/Jan/23 05:03

Description

dt = datetime.date(2021, 12, 27)

# Note; number var in Python gets converted to LongType column;
# this is not supported by the function, so cast to Integer explicitly
df = self.spark.createDataFrame([Row(date=dt, add=2)], "date date, add integer")

self.assertTrue(
    all(
        df.select(
            date_add(df.date, df.add) == datetime.date(2021, 12, 29),
            date_add(df.date, "add") == datetime.date(2021, 12, 29),
            date_add(df.date, 3) == datetime.date(2021, 12, 30),
        ).first()
    )
)

Traceback (most recent call last):
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/tests/test_functions.py", line 391, in test_date_add_function
    ).first()
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/dataframe.py", line 246, in first
    return self.head()
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/dataframe.py", line 310, in head
    rs = self.head(1)
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/dataframe.py", line 312, in head
    return self.take(n)
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/dataframe.py", line 317, in take
    return self.limit(num).collect()
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/dataframe.py", line 1076, in collect
    table = self._session.client.to_table(query)
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/client.py", line 414, in to_table
    table, _ = self._execute_and_fetch(req)
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/client.py", line 586, in _execute_and_fetch
    self._handle_error(rpc_error)
  File "/Users/s.singh/personal/spark-oss/python/pyspark/sql/connect/client.py", line 625, in _handle_error
    raise SparkConnectAnalysisException(
pyspark.sql.connect.client.SparkConnectAnalysisException: [DATATYPE_MISMATCH.UNEXPECTED_INPUT_TYPE] Cannot resolve "date_add(date, add)" due to data type mismatch: Parameter 2 requires the ("INT" or "SMALLINT" or "TINYINT") type, however "add" has the type "BIGINT".
Plan: 'GlobalLimit 1
+- 'LocalLimit 1
   +- 'Project [unresolvedalias('`==`(date_add(date#753, add#754L), 2021-12-29), None), unresolvedalias('`==`(date_add(date#753, add#754L), 2021-12-29), None), (date_add(date#753, 3) = 2021-12-30) AS (date_add(date, 3) = DATE '2021-12-30')#759]
      +- Project [date#753, add#754L]
         +- Project [date#749 AS date#753, add#750L AS add#754L]
            +- LocalRelation [date#749, add#750L]

Issue Links

• links to: [Github] Pull Request #39452 (zhengruifeng)
• links to: [Github] Pull Request #39452 (zhengruifeng)