sqlframe 3.22.0__tar.gz → 3.22.1__tar.gz

{sqlframe-3.22.0 → sqlframe-3.22.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.22.0
+Version: 3.22.1
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.22.0 → sqlframe-3.22.1}/setup.py

@@ -40,7 +40,7 @@ setup(
             "pyspark>=2,<3.6",
             "pytest>=8.2.0,<8.4",
             "pytest-forked",
-            "pytest-postgresql>=6,<7",
+            "pytest-postgresql>=6,<8",
             "pytest-xdist>=3.6,<3.7",
             "pre-commit>=3.7,<5",
             "ruff>=0.4.4,<0.10",

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe/_version.py

@@ -1,8 +1,13 @@
-# file generated by setuptools_scm
+# file generated by setuptools-scm
 # don't change, don't track in version control
+
+__all__ = ["__version__", "__version_tuple__", "version", "version_tuple"]
+
 TYPE_CHECKING = False
 if TYPE_CHECKING:
-    from typing import Tuple, Union
+    from typing import Tuple
+    from typing import Union
+
     VERSION_TUPLE = Tuple[Union[int, str], ...]
 else:
     VERSION_TUPLE = object
@@ -12,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.22.0'
-__version_tuple__ = version_tuple = (3, 22, 0)
+__version__ = version = '3.22.1'
+__version_tuple__ = version_tuple = (3, 22, 1)

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe/base/catalog.py

@@ -70,13 +70,16 @@ class _BaseCatalog(t.Generic[SESSION, DF]):
         }
 
     def add_table(
-        self, table: exp.Table | str, column_mapping: t.Optional[ColumnMapping] = None
+        self,
+        table: exp.Table | str,
+        column_mapping: t.Optional[ColumnMapping] = None,
+        **kwargs: t.Any,
     ) -> None:
         # TODO: Making this an update or add
         table = self.ensure_table(table)
         if self._schema.find(table):
             return
-        if not column_mapping:
+        if column_mapping is None:
             try:
                 column_mapping = {
                     normalize_string(
@@ -100,7 +103,7 @@ class _BaseCatalog(t.Generic[SESSION, DF]):
             if column.this.quoted:
                 self._quoted_columns[table].append(column.this.name)
 
-        self._schema.add_table(table, column_mapping, dialect=self.session.input_dialect)
+        self._schema.add_table(table, column_mapping, dialect=self.session.input_dialect, **kwargs)
 
     def getDatabase(self, dbName: str) -> Database:
         """Get the database with the specified name.

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe/base/util.py

@@ -279,15 +279,6 @@ def verify_openai_installed():
         )
 
 
-def verify_numpy_installed():
-    try:
-        import numpy  # noqa
-    except ImportError:
-        raise ImportError(
-            """Numpy is required for this functionality. `pip install "sqlframe[pandas]"` (also include your engine if needed) to install pandas/numpy."""
-        )
-
-
 def quote_preserving_alias_or_name(col: t.Union[exp.Column, exp.Alias]) -> str:
     from sqlframe.base.session import _BaseSession
 

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe/base/window.py

@@ -82,37 +82,26 @@ class WindowSpec:
     def _calc_start_end(
         self, start: int, end: int
     ) -> t.Dict[str, t.Optional[t.Union[str, exp.Expression]]]:
-        kwargs: t.Dict[str, t.Optional[t.Union[str, exp.Expression]]] = {
-            "start_side": None,
-            "end_side": None,
+        def get_value_and_side(x: int) -> t.Tuple[t.Union[str, exp.Expression], t.Optional[str]]:
+            if x == Window.currentRow:
+                return "CURRENT ROW", None
+            if x < 0:
+                side = "PRECEDING"
+                value = "UNBOUNDED" if x <= Window.unboundedPreceding else F.lit(abs(x)).expression
+                return value, side
+            else:
+                side = "FOLLOWING"
+                value = "UNBOUNDED" if x >= Window.unboundedFollowing else F.lit(x).expression
+                return value, side
+
+        start, start_side = get_value_and_side(start)  # type: ignore
+        end, end_side = get_value_and_side(end)  # type: ignore
+        return {
+            "start": start,  # type: ignore
+            "start_side": start_side,
+            "end": end,  # type: ignore
+            "end_side": end_side,
         }
-        if start == Window.currentRow:
-            kwargs["start"] = "CURRENT ROW"
-        else:
-            kwargs = {
-                **kwargs,
-                **{
-                    "start_side": "PRECEDING",
-                    "start": (
-                        "UNBOUNDED"
-                        if start <= Window.unboundedPreceding
-                        else F.lit(start).expression
-                    ),
-                },
-            }
-        if end == Window.currentRow:
-            kwargs["end"] = "CURRENT ROW"
-        else:
-            kwargs = {
-                **kwargs,
-                **{
-                    "end_side": "FOLLOWING",
-                    "end": (
-                        "UNBOUNDED" if end >= Window.unboundedFollowing else F.lit(end).expression
-                    ),
-                },
-            }
-        return kwargs
 
     def rowsBetween(self, start: int, end: int) -> WindowSpec:
         window_spec = self.copy()

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe/duckdb/session.py

@@ -4,7 +4,7 @@ import typing as t
 from functools import cached_property
 
 from sqlframe.base.session import _BaseSession
-from sqlframe.base.util import soundex, verify_numpy_installed
+from sqlframe.base.util import soundex
 from sqlframe.duckdb.catalog import DuckDBCatalog
 from sqlframe.duckdb.dataframe import DuckDBDataFrame
 from sqlframe.duckdb.readwriter import (
@@ -46,8 +46,6 @@ class DuckDBSession(
         if not hasattr(self, "_conn"):
             conn = conn or duckdb.connect()
             try:
-                # Creating a function requires numpy to be installed so if they don't have it, we'll just skip it
-                verify_numpy_installed()
                 conn.create_function("SOUNDEX", lambda x: soundex(x), return_type=VARCHAR)
             except ImportError:
                 pass

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.22.0
+Version: 3.22.1
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.22.0 → sqlframe-3.22.1}/sqlframe.egg-info/requires.txt

@@ -21,7 +21,7 @@ psycopg<4,>=3.1
 pyarrow<20,>=10
 pyspark<3.6,>=2
 pytest-forked
-pytest-postgresql<7,>=6
+pytest-postgresql<8,>=6
 pytest-xdist<3.7,>=3.6
 pytest<8.4,>=8.2.0
 ruff<0.10,>=0.4.4

{sqlframe-3.22.0 → sqlframe-3.22.1}/tests/integration/test_int_dataframe.py

@@ -6,8 +6,10 @@ import pandas as pd
 import pytest
 from _pytest.fixtures import FixtureRequest
 from pyspark.sql import DataFrame as PySparkDataFrame
+from pyspark.sql import Window
 from pyspark.sql import functions as F
 
+from sqlframe.standalone import Window as SWindow
 from sqlframe.standalone import functions as SF
 from sqlframe.standalone.dataframe import StandaloneDataFrame
 from tests.integration.fixtures import StandaloneSession, is_snowflake
@@ -2548,3 +2550,87 @@ def test_full_outer_nulls_no_match(
     dfs = dfs_concept_1.join(dfs_concept_2, on="col1", how="outer")
 
     compare_frames(df, dfs, compare_schema=False)
+
+
+# https://github.com/eakmanrq/sqlframe/issues/317
+def test_rows_between_negative_start(
+    pyspark_employee: PySparkDataFrame,
+    get_df: t.Callable[[str], BaseDataFrame],
+    compare_frames: t.Callable,
+):
+    session = pyspark_employee.sparkSession
+    df = session.createDataFrame(
+        pd.DataFrame(
+            {"store": [1, 1, 1, 2, 2, 2, 2], "price": [1, 2, 3, 4, 3, 2, 1], "date": list(range(7))}
+        )
+    )
+    window = Window().partitionBy("store").orderBy("date").rowsBetween(-1, 1)
+    df = df.withColumn("rolling_price", F.mean("price").over(window))
+
+    employee = get_df("employee")
+    sf_session = employee.session
+    dfs = sf_session.createDataFrame(
+        pd.DataFrame(
+            {"store": [1, 1, 1, 2, 2, 2, 2], "price": [1, 2, 3, 4, 3, 2, 1], "date": list(range(7))}
+        )
+    )
+    swindow = SWindow().partitionBy("store").orderBy("date").rowsBetween(-1, 1)
+    dfs = dfs.withColumn("rolling_price", SF.mean("price").over(swindow))
+
+    compare_frames(df, dfs, compare_schema=False)
+
+
+# https://github.com/eakmanrq/sqlframe/issues/317
+def test_rows_between_negative_end(
+    pyspark_employee: PySparkDataFrame,
+    get_df: t.Callable[[str], BaseDataFrame],
+    compare_frames: t.Callable,
+):
+    session = pyspark_employee.sparkSession
+    df = session.createDataFrame(
+        pd.DataFrame(
+            {"store": [1, 1, 1, 2, 2, 2, 2], "price": [1, 2, 3, 4, 3, 2, 1], "date": list(range(7))}
+        )
+    )
+    window = Window().partitionBy("store").orderBy("date").rowsBetween(-2, -1)
+    df = df.withColumn("rolling_price", F.mean("price").over(window))
+
+    employee = get_df("employee")
+    sf_session = employee.session
+    dfs = sf_session.createDataFrame(
+        pd.DataFrame(
+            {"store": [1, 1, 1, 2, 2, 2, 2], "price": [1, 2, 3, 4, 3, 2, 1], "date": list(range(7))}
+        )
+    )
+    swindow = SWindow().partitionBy("store").orderBy("date").rowsBetween(-2, -1)
+    dfs = dfs.withColumn("rolling_price", SF.mean("price").over(swindow))
+
+    compare_frames(df, dfs, compare_schema=False)
+
+
+# https://github.com/eakmanrq/sqlframe/issues/317
+def test_rows_between_positive_both_start_end(
+    pyspark_employee: PySparkDataFrame,
+    get_df: t.Callable[[str], BaseDataFrame],
+    compare_frames: t.Callable,
+):
+    session = pyspark_employee.sparkSession
+    df = session.createDataFrame(
+        pd.DataFrame(
+            {"store": [1, 1, 1, 2, 2, 2, 2], "price": [1, 2, 3, 4, 3, 2, 1], "date": list(range(7))}
+        )
+    )
+    window = Window().partitionBy("store").orderBy("date").rowsBetween(1, 2)
+    df = df.withColumn("rolling_price", F.mean("price").over(window))
+
+    employee = get_df("employee")
+    sf_session = employee.session
+    dfs = sf_session.createDataFrame(
+        pd.DataFrame(
+            {"store": [1, 1, 1, 2, 2, 2, 2], "price": [1, 2, 3, 4, 3, 2, 1], "date": list(range(7))}
+        )
+    )
+    swindow = SWindow().partitionBy("store").orderBy("date").rowsBetween(1, 2)
+    dfs = dfs.withColumn("rolling_price", SF.mean("price").over(swindow))
+
+    compare_frames(df, dfs, compare_schema=False)

{sqlframe-3.22.0 → sqlframe-3.22.1}/tests/unit/standalone/test_column.py

@@ -207,7 +207,9 @@ def test_over():
     assert (
         (
             F.sum("cola").over(
-                Window.partitionBy("colb").orderBy("colc").rowsBetween(1, Window.unboundedFollowing)
+                Window.partitionBy("colb")
+                .orderBy("colc")
+                .rowsBetween(-1, Window.unboundedFollowing)
             )
         ).sql()
         == "SUM(cola) OVER (PARTITION BY colb ORDER BY colc ROWS BETWEEN 1 PRECEDING AND UNBOUNDED FOLLOWING)"
@@ -217,7 +219,7 @@ def test_over():
             F.sum("cola").over(
                 Window.partitionBy("colb")
                 .orderBy("colc")
-                .rangeBetween(1, Window.unboundedFollowing)
+                .rangeBetween(-1, Window.unboundedFollowing)
             )
         ).sql()
         == "SUM(cola) OVER (PARTITION BY colb ORDER BY colc RANGE BETWEEN 1 PRECEDING AND UNBOUNDED FOLLOWING)"

{sqlframe-3.22.0 → sqlframe-3.22.1}/tests/unit/standalone/test_window.py

@@ -11,29 +11,41 @@ from sqlframe.standalone.window import Window, WindowSpec
         (Window.partitionBy(F.col("cola"), F.col("colb")), "OVER (PARTITION BY cola, colb)"),
         (WindowSpec().orderBy("cola", "colb"), "OVER (ORDER BY cola, colb)"),
         (Window.orderBy("cola", "colb"), "OVER (ORDER BY cola, colb)"),
-        (WindowSpec().rowsBetween(3, 5), "OVER (ROWS BETWEEN 3 PRECEDING AND 5 FOLLOWING)"),
-        (Window.rowsBetween(3, 5), "OVER (ROWS BETWEEN 3 PRECEDING AND 5 FOLLOWING)"),
-        (WindowSpec().rangeBetween(3, 5), "OVER (RANGE BETWEEN 3 PRECEDING AND 5 FOLLOWING)"),
-        (Window.rangeBetween(3, 5), "OVER (RANGE BETWEEN 3 PRECEDING AND 5 FOLLOWING)"),
+        (WindowSpec().rowsBetween(3, 5), "OVER (ROWS BETWEEN 3 FOLLOWING AND 5 FOLLOWING)"),
+        (Window.rowsBetween(3, 5), "OVER (ROWS BETWEEN 3 FOLLOWING AND 5 FOLLOWING)"),
+        (WindowSpec().rangeBetween(3, 5), "OVER (RANGE BETWEEN 3 FOLLOWING AND 5 FOLLOWING)"),
+        (Window.rangeBetween(3, 5), "OVER (RANGE BETWEEN 3 FOLLOWING AND 5 FOLLOWING)"),
         (
             Window.rowsBetween(Window.unboundedPreceding, 2),
             "OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND 2 FOLLOWING)",
         ),
         (
             Window.rowsBetween(1, Window.unboundedFollowing),
-            "OVER (ROWS BETWEEN 1 PRECEDING AND UNBOUNDED FOLLOWING)",
+            "OVER (ROWS BETWEEN 1 FOLLOWING AND UNBOUNDED FOLLOWING)",
         ),
         (
             Window.rowsBetween(Window.unboundedPreceding, Window.unboundedFollowing),
             "OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING)",
         ),
+        (
+            Window.rowsBetween(-1, Window.currentRow),
+            "OVER (ROWS BETWEEN 1 PRECEDING AND CURRENT ROW)",
+        ),
+        (
+            Window.rowsBetween(-1, 2),
+            "OVER (ROWS BETWEEN 1 PRECEDING AND 2 FOLLOWING)",
+        ),
+        (
+            Window.rowsBetween(-2, -1),
+            "OVER (ROWS BETWEEN 2 PRECEDING AND 1 PRECEDING)",
+        ),
         (
             Window.rangeBetween(Window.unboundedPreceding, 2),
             "OVER (RANGE BETWEEN UNBOUNDED PRECEDING AND 2 FOLLOWING)",
         ),
         (
             Window.rangeBetween(1, Window.unboundedFollowing),
-            "OVER (RANGE BETWEEN 1 PRECEDING AND UNBOUNDED FOLLOWING)",
+            "OVER (RANGE BETWEEN 1 FOLLOWING AND UNBOUNDED FOLLOWING)",
         ),
         (
             Window.rangeBetween(Window.unboundedPreceding, Window.unboundedFollowing),