Add ability to read CSV without header row

pydax/loaders/_table.py

@@ -37,6 +37,7 @@ def load(self, path: Union[_typing.PathLike, Dict[str, str]], options: SchemaDic
                - ``columns`` key specifies the data type of each column. Each data type corresponds to a Pandas'
                  supported dtype. If unspecified, then it is default.
                - ``delimiter`` key specifies the delimiter of the input CSV file.
+               - ``header`` key specifies if the first row of the CSV file contains the headers. Defaults to True
                - ``encoding`` key specifies the encoding of the CSV file. Defaults to UTF-8.
         :raises TypeError: ``path`` is not a path object.
         """
@@ -55,9 +56,15 @@ def load(self, path: Union[_typing.PathLike, Dict[str, str]], options: SchemaDic
             else:
                 dtypes[column] = type_
 
+        names = None
+        if options.get('header', True) is False:
+            # If no header use the columns provided in schema
+            names = [*options.get('columns', {})]
+
         return pd.read_csv(path, dtype=dtypes,
                            # The following line after "if" is for circumventing
                            # https://github.com/pandas-dev/pandas/issues/38489
                            parse_dates=parse_dates if len(parse_dates) > 0 else False,
+                           names=names,
                            encoding=options.get('encoding', 'utf-8'),
                            delimiter=options.get('delimiter', ','))

tests/test_loaders.py

@@ -243,3 +243,13 @@ def test_csv_pandas_loader_no_encoding(self, tmp_path, noaa_jfk_schema):
 
         del noaa_jfk_schema['subdatasets']['jfk_weather_cleaned']['format']['options']['encoding']
         self.test_csv_pandas_loader(tmp_path, noaa_jfk_schema)
+
+    def test_csv_pandas_header(self, tmp_path, noaa_jfk_schema):
+        "Test CSVPandasLoader header options"
+
+        noaa_jfk_schema['subdatasets']['jfk_weather_cleaned']['format']['options']['header'] = True
+        self.test_csv_pandas_loader(tmp_path, noaa_jfk_schema)
+
+        with pytest.raises(ValueError):  # Pandas should error from trying to read string as another dtype
+            noaa_jfk_schema['subdatasets']['jfk_weather_cleaned']['format']['options']['header'] = False
+            Dataset(noaa_jfk_schema, tmp_path, mode=Dataset.InitializationMode.DOWNLOAD_AND_LOAD)