UNABLE_TO_INFER_SCHEMA on pyspark

[cometta]

i get below error when convert from sparkml model to ONNX

An error was encountered:
AnalysisException
[Traceback (most recent call last):
,   File "/tmp/spark-591fcd26-f35c-4194-9d93-9e4fa0b7a634/shell_wrapper.py", line 113, in exec
    self._exec_then_eval(code)
,   File "/tmp/spark-591fcd26-f35c-4194-9d93-9e4fa0b7a634/shell_wrapper.py", line 106, in _exec_then_eval
    exec(compile(last, '<string>', 'single'), self.globals)
,   File "<string>", line 1, in <module>
,   File "/home/user/work/.python_libs/lib/python3.10/site-packages/onnxmltools/convert/main.py", line 302, in convert_sparkml
    return convert(
,   File "/home/user/work/.python_libs/lib/python3.10/site-packages/onnxmltools/convert/sparkml/convert.py", line 101, in convert
    onnx_model = convert_topology(
,   File "/home/user/work/.python_libs/lib/python3.10/site-packages/onnxconverter_common/topology.py", line 776, in convert_topology
    get_converter(operator.type)(scope, operator, container)
,   File "/home/user/work/.python_libs/lib/python3.10/site-packages/onnxmltools/convert/sparkml/operator_converters/random_forest_regressor.py", line 31, in convert_random_forest_regressor
    tree_df = save_read_sparkml_model_data(
,   File "/home/user/work/.python_libs/lib/python3.10/site-packages/onnxmltools/convert/sparkml/operator_converters/tree_ensemble_common.py", line 113, in save_read_sparkml_model_data
    df = spark.read.parquet(os.path.join(path, "data"))
,   File "/opt/spark/python/lib/pyspark.zip/pyspark/sql/readwriter.py", line 531, in parquet
    return self._df(self._jreader.parquet(_to_seq(self._spark._sc, paths)))
,   File "/opt/spark/python/lib/py4j-0.10.9.7-src.zip/py4j/java_gateway.py", line 1322, in __call__
    return_value = get_return_value(
,   File "/opt/spark/python/lib/pyspark.zip/pyspark/errors/exceptions/captured.py", line 175, in deco
    raise converted from None
, pyspark.errors.exceptions.captured.AnalysisException: [UNABLE_TO_INFER_SCHEMA] Unable to infer schema for Parquet. It must be specified manually.
]

print(initial_types)

[('some_column_name', StringTensorType(shape=[None, 1])), ('var1', StringTensorType(shape=[None, 1])), ('var2', FloatTensorType(shape=[None, 1])), ('var3', FloatTensorType(shape=[None, 1])), ('var4', FloatTensorType(shape=[None, 1])), ('var5', FloatTensorType(shape=[None, 1])), ('var6', FloatTensorType(shape=[None, 1])), ('var7', FloatTensorType(shape=[None, 1])), ('var9', FloatTensorType(shape=[None, 1]))]

model i use is from pipeline randomforest

onnx_model = convert_sparkml(model, 'pyspark test', initial_types, spark_session = spark)