lightweight-table-diff 0.1.1__tar.gz → 0.1.3__tar.gz

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lightweight-table-diff
-Version: 0.1.1
+Version: 0.1.3
 Summary: Cell-level table diffing for Polars
 Requires-Python: >=3.11.1
 Description-Content-Type: text/markdown

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "lightweight-table-diff"
-version = "0.1.1"
+version = "0.1.3"
 description = "Cell-level table diffing for Polars"
 readme = "README.md"
 requires-python = ">=3.11.1"

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/src/lightweight_table_diff/adapters/__init__.py

@@ -25,7 +25,9 @@ def load(source_def: dict, **context: Any) -> pl.LazyFrame:
 from .csv import load_csv      # noqa: E402
 from .parquet import load_parquet  # noqa: E402
 from .sav import load_sav      # noqa: E402
+from .hive_s3 import load_hive # noqa: E402
 
 register("parquet", load_parquet)
 register("csv", load_csv)
-register("sav", load_sav)
+register("sav", load_sav)
+register("hive", load_hive)

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/src/lightweight_table_diff/adapters/sav.py

@@ -11,7 +11,7 @@ def load_sav(source_def: dict) -> pl.LazyFrame:
         import pyreadstat
     except ImportError:
         raise ImportError(
-            "pyreadstat is required for .sav files: pip install pyreadstat"
+            "pyreadstat is required for .sav files: pip install 'lightweight-table-diff[spss]'"
         ) from None
 
     pdf, _meta = pyreadstat.read_sav(

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/src/lightweight_table_diff/core.py

@@ -20,7 +20,7 @@ def diff_tbls(
     after: pl.LazyFrame,
     id_cols: list[str],
     compare_cols: list[str] | None = None,
-    join_type: str = "full",
+    join_type: str = "inner",
 ) -> pl.LazyFrame:
     if compare_cols is None:
         compare_cols = list(before.drop(*id_cols).collect_schema().keys())
@@ -69,7 +69,7 @@ def batch_diff_tbls(
     id_cols: list[str],
     compare_cols: list[str] | None = None,
     batch_size: int = 50,
-    join_type: str = "full",
+    join_type: str = "inner",
 ) -> pl.LazyFrame:
     if compare_cols is None:
         compare_cols = list(before.drop(*id_cols).collect_schema().keys())

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/src/lightweight_table_diff/runner.py

@@ -72,7 +72,7 @@ def run_comparison(job: dict, **context) -> ComparisonResult:
         keys,
         cols,
         batch_size=job.get("batch_size", 50),
-        join_type=job.get("join_type", "full"),
+        join_type="inner",
     )
 
     n = diff.select(pl.len()).collect().item()

{lightweight_table_diff-0.1.1 → lightweight_table_diff-0.1.3}/src/lightweight_table_diff.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lightweight-table-diff
-Version: 0.1.1
+Version: 0.1.3
 Summary: Cell-level table diffing for Polars
 Requires-Python: >=3.11.1
 Description-Content-Type: text/markdown