Datasets:
thisisanshgupta
/
Pycode

Dataset card Data Studio Files
xet
Community
text
stringlengths
0
828
one_col_df = pd.read_csv(file_path, encoding=encoding, **kwargs)
if one_col_df.shape[1] == 1:
return one_col_df[one_col_df.columns[0]]
else:
raise Exception('Cannot build a series from this csv: it has more than two columns (one index + one value).'
' Probably the parsing chain $read_df_or_series_from_csv => single_row_or_col_df_to_series$'
'will work, though.')
else:
return pd.read_csv(file_path, encoding=encoding, **kwargs)"
1533,"def get_default_pandas_parsers() -> List[AnyParser]:
""""""
Utility method to return the default parsers able to parse a dictionary from a file.
:return:
""""""
return [SingleFileParserFunction(parser_function=read_dataframe_from_xls,
streaming_mode=False,
supported_exts={'.xls', '.xlsx', '.xlsm'},
supported_types={pd.DataFrame},
option_hints=pandas_parsers_option_hints_xls),
SingleFileParserFunction(parser_function=read_df_or_series_from_csv,
streaming_mode=False,
supported_exts={'.csv', '.txt'},
supported_types={pd.DataFrame, pd.Series},
option_hints=pandas_parsers_option_hints_csv),
]"
1534,"def dict_to_df(desired_type: Type[T], dict_obj: Dict, logger: Logger, orient: str = None, **kwargs) -> pd.DataFrame:
""""""
Helper method to convert a dictionary into a dataframe. It supports both simple key-value dicts as well as true
table dicts. For this it uses pd.DataFrame constructor or pd.DataFrame.from_dict intelligently depending on the
case.
The orientation of the resulting dataframe can be configured, or left to default behaviour. Default orientation is
different depending on the contents:
* 'index' for 2-level dictionaries, in order to align as much as possible with the natural way to express rows in
JSON
* 'columns' for 1-level (simple key-value) dictionaries, so as to preserve the data types of the scalar values in
the resulting dataframe columns if they are different
:param desired_type:
:param dict_obj:
:param logger:
:param orient: the orientation of the resulting dataframe.
:param kwargs:
:return:
""""""
if len(dict_obj) > 0:
first_val = dict_obj[next(iter(dict_obj))]
if isinstance(first_val, dict) or isinstance(first_val, list):
# --'full' table
# default is index orientation
orient = orient or 'index'
# if orient is 'columns':
# return pd.DataFrame(dict_obj)
# else:
return pd.DataFrame.from_dict(dict_obj, orient=orient)
else:
# --scalar > single-row or single-col
# default is columns orientation
orient = orient or 'columns'
if orient is 'columns':
return pd.DataFrame(dict_obj, index=[0])
else:
res = pd.DataFrame.from_dict(dict_obj, orient=orient)
res.index.name = 'key'
return res.rename(columns={0: 'value'})
else:
# for empty dictionaries, orientation does not matter
# but maybe we should still create a column 'value' in this empty dataframe ?
return pd.DataFrame.from_dict(dict_obj)"
1535,"def single_row_or_col_df_to_series(desired_type: Type[T], single_rowcol_df: pd.DataFrame, logger: Logger, **kwargs)\
-> pd.Series:
""""""
Helper method to convert a dataframe with one row or one or two columns into a Series
:param desired_type:
:param single_col_df:
:param logger:
:param kwargs:
:return:
""""""
if single_rowcol_df.shape[0] == 1:
# one row
return single_rowcol_df.transpose()[0]
elif single_rowcol_df.shape[1] == 2 and isinstance(single_rowcol_df.index, pd.RangeIndex):
# two columns but the index contains nothing but the row number : we can use the first column
d = single_rowcol_df.set_index(single_rowcol_df.columns[0])
return d[d.columns[0]]
elif single_rowcol_df.shape[1] == 1:
# one column and one index
d = single_rowcol_df
return d[d.columns[0]]
else:
raise ValueError('Unable to convert provided dataframe to a series : '
'expected exactly 1 row or 1 column, found : ' + str(single_rowcol_df.shape) + '')"