Given a list of names, types, and optionally values, construct a Schema.
(
col_names, col_types=None,
col_blobs=None, col_metadata=None
)
| 1059 | |
| 1060 | |
| 1061 | def from_column_list( |
| 1062 | col_names, col_types=None, |
| 1063 | col_blobs=None, col_metadata=None |
| 1064 | ): |
| 1065 | """ |
| 1066 | Given a list of names, types, and optionally values, construct a Schema. |
| 1067 | """ |
| 1068 | if col_types is None: |
| 1069 | col_types = [None] * len(col_names) |
| 1070 | if col_metadata is None: |
| 1071 | col_metadata = [None] * len(col_names) |
| 1072 | if col_blobs is None: |
| 1073 | col_blobs = [None] * len(col_names) |
| 1074 | assert len(col_names) == len(col_types), ( |
| 1075 | 'col_names and col_types must have the same length.' |
| 1076 | ) |
| 1077 | assert len(col_names) == len(col_metadata), ( |
| 1078 | 'col_names and col_metadata must have the same length.' |
| 1079 | ) |
| 1080 | assert len(col_names) == len(col_blobs), ( |
| 1081 | 'col_names and col_blobs must have the same length.' |
| 1082 | ) |
| 1083 | root = _SchemaNode('root', 'Struct') |
| 1084 | for col_name, col_type, col_blob, col_md in zip( |
| 1085 | col_names, col_types, col_blobs, col_metadata |
| 1086 | ): |
| 1087 | columns = col_name.split(FIELD_SEPARATOR) |
| 1088 | current = root |
| 1089 | for i in range(len(columns)): |
| 1090 | name = columns[i] |
| 1091 | type_str = '' |
| 1092 | field = None |
| 1093 | if i == len(columns) - 1: |
| 1094 | type_str = col_type |
| 1095 | field = Scalar( |
| 1096 | dtype=col_type, |
| 1097 | blob=col_blob, |
| 1098 | metadata=col_md |
| 1099 | ) |
| 1100 | next = current.add_child(name, type_str) |
| 1101 | if field is not None: |
| 1102 | next.field = field |
| 1103 | current = next |
| 1104 | |
| 1105 | return root.get_field() |
| 1106 | |
| 1107 | |
| 1108 | def from_blob_list(schema, values, throw_on_type_mismatch=False): |