def _first_valid_object(data_frame, column_name):
"""
Returns the first object in a pandas DataFrame column that does not represent a missing type (None, NaN or NaT).
"""
for cell in data_frame[column_name]:
if not is_missing(cell):
return cell
return None
def _scan_column_for_type(data_frame, column_name):
"""
Get the type of a column (fails if multiple types are found). Numpy scalar types are converted to dtypes.
"""
col_type = None
for cell in data_frame[column_name]:
if not is_missing(cell):
if col_type is not None:
if not types_are_equivalent(type(cell), col_type):
raise ValueError('More than one type in column ' + str(column_name) + '. Found '
+ col_type.__name__ + ' and ' + type(cell).__name__)
else:
col_type = type(cell)
col_type = Serializer._convert_to_dtype_if_numpy_type(col_type)
return col_type
def _scan_list_column_for_element_type(data_frame, column_name):
"""
Get the type of a list column (fails if multiple types are found). Numpy scalar types are converted to dtypes.
"""
col_type = None
for list_cell in data_frame[column_name]:
if list_cell is not None:
for cell in list_cell:
if not is_missing(cell):
if col_type is not None:
if not types_are_equivalent(type(cell), col_type):
raise ValueError('More than one type in column ' + str(column_name) + '. Found '
+ col_type.__name__ + ' and ' + type(cell).__name__)
else:
col_type = type(cell)
if col_type == list or col_type == set:
raise ValueError('Output table contains a nested collection. Nested collections are not yet supported.')
col_type = Serializer._convert_to_dtype_if_numpy_type(col_type)
return col_type
