pydiverse-common 0.3.7__tar.gz → 0.3.8__tar.gz

{pydiverse_common-0.3.7 → pydiverse_common-0.3.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pydiverse-common
-Version: 0.3.7
+Version: 0.3.8
 Summary: Common functionality shared between pydiverse libraries
 Author: QuantCo, Inc.
 Author-email: Martin Trautmann <windiana@users.sf.net>, Finn Rudolph <finn.rudolph@t-online.de>

{pydiverse_common-0.3.7 → pydiverse_common-0.3.8}/docs/source/changelog.md

@@ -1,5 +1,8 @@
 # Changelog
 
+## 0.3.8 (2025-08-19)
+- fixed util.hashing.hash_polars_dataframe for simple dataframe
+
 ## 0.3.7 (2025-08-19)
 - support util.hashing.hash_polars_dataframe
 

{pydiverse_common-0.3.7 → pydiverse_common-0.3.8}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pydiverse-common"
-version = "0.3.7"
+version = "0.3.8"
 description = "Common functionality shared between pydiverse libraries"
 authors = [
   { name = "QuantCo, Inc." },

{pydiverse_common-0.3.7 → pydiverse_common-0.3.8}/src/pydiverse/common/util/hashing.py

@@ -73,24 +73,27 @@ def hash_polars_dataframe(df: pl.DataFrame, use_init_repr=False) -> str:
                 list_columns = [
                     col for col, dtype in df.schema.items() if dtype == pl.List
                 ]
-                content_hash = str(
-                    df.lazy()
-                    .with_columns(pl.col(array_columns).reshape([-1]).implode())
-                    .with_columns(
-                        # Necessary because hash() does not work on lists of strings.
-                        # This can be removed when
-                        # https://github.com/pola-rs/polars/issues/21523 is resolved
-                        # in all supported versions of polars.
-                        pl.selectors.by_dtype(pl.List(pl.String)).list.eval(
-                            pl.element().hash()
-                        )
+                lf = df.lazy()
+                if array_columns:
+                    lf = lf.with_columns(pl.col(array_columns).reshape([-1]).implode())
+                lf = lf.with_columns(
+                    # Necessary because hash() does not work on lists of strings.
+                    # This can be removed when
+                    # https://github.com/pola-rs/polars/issues/21523 is resolved
+                    # in all supported versions of polars.
+                    pl.selectors.by_dtype(pl.List(pl.String)).list.eval(
+                        pl.element().hash()
                     )
+                )
+                if list_columns or array_columns:
                     # Necessary because hash_rows() does not work on lists.
                     # This can be removed when
                     # https://github.com/pola-rs/polars/issues/24121 is resolved
                     # in all supported versions of polars.
-                    .with_columns(pl.col(*list_columns, *array_columns).hash())
-                    .collect()
+                    lf = lf.with_columns(pl.col(*list_columns, *array_columns).hash())
+
+                content_hash = str(
+                    lf.collect()
                     .hash_rows()  # We get a Series of hashes, one for each row
                     # Since polars only hashes rows, we need to implode the Series into
                     # a single row to get a single hash

{pydiverse_common-0.3.7 → pydiverse_common-0.3.8}/tests/test_util.py

@@ -142,6 +142,17 @@ def check_df_hashes(df1_a: pl.DataFrame, other_dfs: list[pl.DataFrame]) -> None:
         ) == hash_polars_dataframe(df1_other, use_init_repr=True)
 
 
+@pytest.mark.skipif(pl.DataFrame is None, reason="requires polars")
+def test_hashing_basic():
+    df1_a = pl.DataFrame(dict(x=[1]))
+    df1_b = pl.DataFrame(dict(y=[1]))
+    df1_c = pl.DataFrame(dict(x=[2]))
+    df1_d = pl.DataFrame(dict(x=[1.0]))
+    df1_e = pl.DataFrame(dict(x=[]))
+
+    check_df_hashes(df1_a, [df1_b, df1_c, df1_d, df1_e])
+
+
 @pytest.mark.skipif(pl.DataFrame is None, reason="requires polars")
 def test_hashing():
     df1_a = pl.DataFrame(