🐝 Update pandas to 2.1

etl/datadiff.py

@@ -325,10 +325,10 @@ def cli(
         ds_a = _match_dataset(path_to_ds_a, path)
         ds_b = _match_dataset(path_to_ds_b, path)
 
-        if ds_a and ds_b and ds_a.metadata.source_checksum == ds_b.metadata.source_checksum:
-            # skip if they have the same source checksum, note that we're not comparing checksum of actual data
-            # to improve performance. Source checksum should be enough
-            continue
+        # if ds_a and ds_b and ds_a.metadata.source_checksum == ds_b.metadata.source_checksum:
+        #     # skip if they have the same source checksum, note that we're not comparing checksum of actual data
+        #     # to improve performance. Source checksum should be enough
+        #     continue
 
         lines = []
 
@@ -389,8 +389,8 @@ def _index_equals(table_a: pd.DataFrame, table_b: pd.DataFrame, sample: int = 10
 
 def _dict_diff(dict_a: Dict[str, Any], dict_b: Dict[str, Any], tabs: int = 0, **kwargs) -> str:
     """Convert dictionaries into YAML and compare them using difflib. Return colored diff as a string."""
-    meta_a = yaml_dump(dict_a, **kwargs)
-    meta_b = yaml_dump(dict_b, **kwargs)
+    meta_a = yaml_dump(dict_a, sort_keys=True, **kwargs)
+    meta_b = yaml_dump(dict_b, sort_keys=True, **kwargs)
 
     lines = difflib.ndiff(meta_a.splitlines(keepends=True), meta_b.splitlines(keepends=True))  # type: ignore
     # do not print lines that are identical

etl/files.py

@@ -156,14 +156,15 @@ def yaml_dump(
     strip_lines: bool = True,
     replace_confusing_ascii: bool = False,
     width: int = 120,
+    sort_keys: bool = False,
 ) -> Optional[str]:
     """Alternative to yaml.dump which produces good looking multi-line strings and perserves ordering
     of keys. If strip_lines is True, all lines in the string will be stripped and all tabs will be
     replaced by two spaces."""
     # strip lines, otherwise YAML won't output strings in literal format
     if strip_lines:
         d = _strip_lines_in_dict(d)
-    s = yaml.dump(d, stream=stream, sort_keys=False, allow_unicode=True, Dumper=_MyDumper, width=width)
+    s = yaml.dump(d, stream=stream, sort_keys=sort_keys, allow_unicode=True, Dumper=_MyDumper, width=width)
     if replace_confusing_ascii:
         assert s, "replace_confusing_ascii does not work for streams"
         s = (

lib/catalog/owid/catalog/tables.py

@@ -31,7 +31,6 @@
 import structlog
 from pandas._typing import FilePath, ReadCsvBuffer, Scalar  # type: ignore
 from pandas.core.series import Series
-from pandas.util._decorators import rewrite_axis_style_signature
 
 from owid.repack import repack_frame
 
@@ -425,10 +424,6 @@ def equals_table(self, table: "Table") -> bool:
             and self._fields == table._fields
         )
 
-    @rewrite_axis_style_signature(
-        "mapper",
-        [("copy", True), ("inplace", False), ("level", None), ("errors", "ignore")],
-    )
     def rename(self, *args: Any, **kwargs: Any) -> Optional["Table"]:
         """Rename columns while keeping their metadata."""
         inplace = kwargs.get("inplace")

lib/catalog/pyproject.toml

@@ -11,7 +11,7 @@ homepage = "https://github.com/owid/owid-grapher-py"
 
 [tool.poetry.dependencies]
 python = ">=3.8.1, <3.12"
-pandas = ">=1.3.3,<2.0"
+pandas = ">=2.1.0,<2.2.0"
 jsonschema = ">=3.2.0"
 pyarrow = ">=10.0.1"
 ipdb = ">=0.13.9"