[ARROW-6506] [C++] Validation of ExtensionType with nested type fails

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.15.0
• Component/s: C++
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/22874
• Assignee: Joris Van den Bossche
• Reporter: Joris Van den Bossche
• Created: 10/Sep/19 07:36
• Updated: 11/Jan/23 07:47
• Resolved: 11/Sep/19 10:32

Description

A reproducer using the Python ExtensionType:

class MyStructType(pa.ExtensionType): 

    def __init__(self): 
        storage_type = pa.struct([('a', pa.int64()), ('b', pa.int64())]) 
        pa.ExtensionType.__init__(self, storage_type, 'my_struct_type') 

    def __arrow_ext_serialize__(self): 
        return b'' 

    @classmethod 
    def __arrow_ext_deserialize__(self, storage_type, serialized): 
        return MyStructType() 

ty = MyStructType()
storage_array = pa.array([{'a': 1, 'b': 2}], ty.storage_type) 
arr = pa.ExtensionArray.from_storage(ty, storage_array) 

then validating this array fails because it expects no children (the extension array itself has no children, only the storage array):

In [8]: arr.validate()   
---------------------------------------------------------------------------
ArrowInvalid                              Traceback (most recent call last)
<ipython-input-8-13783ce9f25e> in <module>
----> 1 arr.validate()

~/scipy/repos/arrow/python/pyarrow/array.pxi in pyarrow.lib.Array.validate()

~/scipy/repos/arrow/python/pyarrow/error.pxi in pyarrow.lib.check_status()

ArrowInvalid: Expected 0 child arrays in array of type extension<my_struct_type>, got 2

Issue Links

• links to: GitHub Pull Request #5340