[
https://issues.apache.org/jira/browse/SPARK-53448?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18017399#comment-18017399
]
Vindhya G commented on SPARK-53448:
-----------------------------------
When I ran the code locally I found
https://github.com/apache/spark/blob/master/python/pyspark/sql/pandas/types.py#L1081
has the condition for value to be dict but the above code sends VariantVal.
When I added below check it worked locally.
{code:java}
if isinstance(value, VariantVal):
return value
{code}
> Conversion of a pyspark DataFrame with a Variant column to pandas fails with
> an error
> -------------------------------------------------------------------------------------
>
> Key: SPARK-53448
> URL: https://issues.apache.org/jira/browse/SPARK-53448
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 4.0.0
> Reporter: Vindhya G
> Priority: Minor
>
>
> {code:java}
> {{from pyspark.sql import SparkSessionfrom pyspark.sql.functions import expr
> spark = SparkSession.builder.getOrCreate();
> data = [
> ('\{"name": "Alice", "age": 30, "city": "New York", "salary": 85000}',),
> ('\{"name": "Bob", "age": 25, "city": "San Francisco", "salary":
> 92000}',),
> ('\{"name": "Charlie", "age": 35, "city": "Los Angeles", "salary":
> 79000}',)
> ]
> df = spark.createDataFrame(data, ["json_data"])
> df_variant = df.withColumn("variant_column", expr("parse_json(json_data)"))
> df_variant.printSchema()
> df_variant.show(truncate=False)
> pdf = df_variant.toPandas()print(pdf)}}
> {code}
> The below code throws below error.
> {noformat}
> File
> ~/.pyenv/versions/3.11.9/lib/python3.11/site-packages/pyspark/sql/pandas/types.py:1090,
> in
> _create_converter_to_pandas.<locals>._converter.<locals>.convert_variant(value)
> 1088 return VariantVal(value["value"], value["metadata"])
> 1089 else:
> -> 1090 raise PySparkValueError(errorClass="MALFORMED_VARIANT")
> File
> ~/.pyenv/versions/3.11.9/lib/python3.11/site-packages/pyspark/errors/exceptions/base.py:47,
> in PySparkException.__init__(self, message, errorClass, messageParameters,
> contexts)
> 44 self._error_reader = ErrorClassesReader()
> 46 if message is None:
> ---> 47 self._message = self._error_reader.get_error_message(
> 48 cast(str, errorClass), cast(Dict[str, str], messageParameters)
> 49 )
> 50 else:
> 51 self._message = message
> File
> ~/.pyenv/versions/3.11.9/lib/python3.11/site-packages/pyspark/errors/utils.py:105,
> in ErrorClassesReader.get_error_message(self, errorClass, messageParameters)
> 103 # Verify message parameters.
> 104 message_parameters_from_template = re.findall("<([a-zA-Z0-9_-]+)>",
> message_template)
> --> 105 assert set(message_parameters_from_template) ==
> set(messageParameters), (
> 106 f"Undefined error message parameter for error class:
> {errorClass}. "
> 107 f"Parameters: {messageParameters}"
> 108 )
> 110 def replace_match(match: Match[str]) -> str:
> 111 return match.group().translate(str.maketrans("<>", "{}"))
> TypeError: 'NoneType' object is not iterable
> {noformat}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
