Initial Commit

narwhals-python/README.md

@@ -0,0 +1,16 @@
+# Writing DataFrame-Agnostic Python Code With Narwhals
+
+This folder contains the downloadable code from the Real Python tutorial: [Writing DataFrame-Agnostic Python Code With Narwhals](https://realpython.com/python-narwhals/).
+
+| File                      	| Description                                         	|
+|---------------------------	|-----------------------------------------------------	|
+| `authors.parquet`         	| Data for main sections of the tutorial              	|
+| `books.parquet`           	| Data for the self study section of the tutorial     	|
+| `presidents.parquet`      	| Data for the self study section of the tutorial     	|
+| `exercise_solution.py`    	| A possible solution to the self study exercise      	|
+| `universal_processing.py` 	| Functions used in the main sections of the tutorial 	|
+
+
+
+
+

narwhals-python/exercise_solution.py

@@ -0,0 +1,17 @@
+import narwhals as nw
+from narwhals.typing import IntoFrameT
+
+
+def rowling_books(df: IntoFrameT, lf: IntoFrameT) -> IntoFrameT:
+    return (
+        nw.from_native(df)
+        .join(
+            nw.from_native(lf)
+            .filter(nw.col("last_name").str.contains("Rowling"))
+            .collect(),
+            on="author_id",
+        )
+        .select(["book_title", "year_published", "first_name", "last_name"])
+        .sort("year_published")
+        .to_pandas()
+    )

narwhals-python/universal_processing.py

@@ -0,0 +1,56 @@
+import narwhals as nw
+from narwhals.typing import FrameT, IntoFrameT
+
+
+def universal_groupby_v1(df: IntoFrameT) -> IntoFrameT:
+    return (
+        nw.from_native(df)
+        .group_by("party_name")
+        .agg(nw.col("last_name").count())
+        .sort("party_name")
+        .to_native()
+    )
+
+
+@nw.narwhalify
+def universal_groupby_v2(df: FrameT) -> FrameT:
+    return (
+        df.group_by("party_name")
+        .agg(nw.col("last_name").count())
+        .sort("party_name")
+    )
+
+
+def universal_groupby_v3(df: IntoFrameT) -> IntoFrameT:
+    return (
+        nw.from_native(df)
+        .group_by("party_name")
+        .agg(nw.col("last_name").count())
+        .sort("party_name")
+        .to_polars()
+    )
+
+
+def universal_pivot_v1(df: IntoFrameT) -> IntoFrameT:
+    return (
+        nw.from_native(df)
+        .pivot(
+            on="party_name",
+            index="century",
+            values="last_name",
+            aggregate_function="count",
+        )
+        .to_native()
+    )
+
+
+def universal_pivot_v2(df: IntoFrameT) -> IntoFrameT:
+    df = nw.from_native(df)
+    if isinstance(df, nw.LazyFrame):
+        df = df.collect()
+    return df.pivot(
+        on="party_name",
+        index="century",
+        values="last_name",
+        aggregate_function="count",
+    ).to_native()