sqlframe 3.31.0__py3-none-any.whl → 3.31.2__py3-none-any.whl

sqlframe/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.31.0'
-__version_tuple__ = version_tuple = (3, 31, 0)
+__version__ = version = '3.31.2'
+__version_tuple__ = version_tuple = (3, 31, 2)

sqlframe/base/column.py

@@ -430,8 +430,9 @@ class Column:
         )
 
     def over(self, window: WindowSpec) -> Column:
+        column_expression = self.column_expression.meta.get("window_func", self.column_expression)
         window_expression = window.expression.copy()
-        window_expression.set("this", self.column_expression)
+        window_expression.set("this", column_expression)
         return Column(window_expression)
 
     def getItem(self, key: t.Any) -> Column:

sqlframe/base/functions.py

@@ -625,11 +625,22 @@ def covar_samp(col1: ColumnOrName, col2: ColumnOrName) -> Column:
 def first(col: ColumnOrName, ignorenulls: t.Optional[bool] = None) -> Column:
     session = _get_session()
 
-    if session._is_duckdb:
-        ignorenulls = None
-
     this = Column.invoke_expression_over_column(col, expression.First)
     if ignorenulls:
+        if session._is_duckdb:
+            return Column(
+                expression.Filter(
+                    this=this.expression,
+                    expression=expression.Where(
+                        this=expression.Not(
+                            this=expression.Is(
+                                this=Column.ensure_col(col).expression,
+                                expression=expression.Null(),
+                            )
+                        )
+                    ),
+                )
+            )
         return Column.invoke_expression_over_column(this, expression.IgnoreNulls)
     return this
 
@@ -3266,12 +3277,57 @@ def find_in_set(str: ColumnOrName, str_array: ColumnOrName) -> Column:
     return Column.invoke_anonymous_function(str, "find_in_set", str_array)
 
 
-@meta(unsupported_engines="*")
+@meta(unsupported_engines=["bigquery", "postgres", "snowflake"])
 def first_value(col: ColumnOrName, ignoreNulls: t.Optional[t.Union[bool, Column]] = None) -> Column:
+    """Returns the first value of `col` for a group of rows. It will return the first non-null
+    value it sees when `ignoreNulls` is set to true. If all values are null, then null is returned.
+
+    Parameters
+    ----------
+    col : :class:`~pyspark.sql.Column` or str
+        target column to work on.
+    ignorenulls : :class:`~pyspark.sql.Column` or bool
+        if first value is null then look for first non-null value.
+
+    Returns
+    -------
+    :class:`~pyspark.sql.Column`
+        some value of `col` for a group of rows.
+
+    Examples
+    --------
+    >>> import pyspark.sql.functions as sf
+    >>> spark.createDataFrame(
+    ...     [(None, 1), ("a", 2), ("a", 3), ("b", 8), (None, 2)], ["a", "b"]
+    ... ).select(sf.first_value('a'), sf.first_value('b')).show()
+    +--------------+--------------+
+    |first_value(a)|first_value(b)|
+    +--------------+--------------+
+    |          NULL|             1|
+    +--------------+--------------+
+
+    >>> import pyspark.sql.functions as sf
+    >>> spark.createDataFrame(
+    ...     [(None, 1), ("a", 2), ("a", 3), ("b", 8), (None, 2)], ["a", "b"]
+    ... ).select(sf.first_value('a', True), sf.first_value('b', True)).show()
+    +--------------+--------------+
+    |first_value(a)|first_value(b)|
+    +--------------+--------------+
+    |             a|             1|
+    +--------------+--------------+
+    """
+    session = _get_session()
+
     column = Column.invoke_expression_over_column(col, expression.FirstValue)
 
     if ignoreNulls:
-        return Column(expression.IgnoreNulls(this=column.column_expression))
+        column = Column(expression.IgnoreNulls(this=column.column_expression))
+
+    if session._is_duckdb:
+        agg_func = first(col, ignoreNulls)  # type: ignore
+        agg_func.expression._meta = agg_func.expression._meta or {}
+        agg_func.expression._meta["window_func"] = column.expression
+        return agg_func
     return column
 
 
@@ -3930,13 +3986,16 @@ def last_value(col: ColumnOrName, ignoreNulls: t.Optional[t.Union[bool, Column]]
     """
     session = _get_session()
 
-    if session._is_duckdb:
-        return last(col, ignoreNulls)  # type: ignore
-
     column = Column.invoke_expression_over_column(col, expression.LastValue)
 
     if ignoreNulls:
-        return Column(expression.IgnoreNulls(this=column.column_expression))
+        column = Column(expression.IgnoreNulls(this=column.column_expression))
+
+    if session._is_duckdb:
+        agg_func = last(col, ignoreNulls)  # type: ignore
+        agg_func.expression._meta = agg_func.expression._meta or {}
+        agg_func.expression._meta["window_func"] = column.expression
+        return agg_func
     return column
 
 

{sqlframe-3.31.0.dist-info → sqlframe-3.31.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.31.0
+Version: 3.31.2
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.31.0.dist-info → sqlframe-3.31.2.dist-info}/RECORD

@@ -1,15 +1,15 @@
 sqlframe/__init__.py,sha256=SB80yLTITBXHI2GCDS6n6bN5ObHqgPjfpRPAUwxaots,3403
-sqlframe/_version.py,sha256=AgRbmN0zJrj6Fie24WbIl_HozqASkUgbWo9IOWaS7vU,513
+sqlframe/_version.py,sha256=NVEtY8IR79NwYZzL_LNH-0DstDZDCujwAeO2Mcu8K0E,513
 sqlframe/py.typed,sha256=Nqnn8clbgv-5l0PgxcTOldg8mkMKrFn4TvPL-rYUUGg,1
 sqlframe/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 sqlframe/base/_typing.py,sha256=b2clI5HI1zEZKB_3Msx3FeAJQyft44ubUifJwQRVXyQ,1298
 sqlframe/base/catalog.py,sha256=ZuU_qmt4yjSoTYgecSGnOhitOdh3rJbGCUjnUBp5mlc,38564
-sqlframe/base/column.py,sha256=AG9Z_6RNhVxLhLU29kRCgzMgDNSm-_GFg96xLqk1-bs,19838
+sqlframe/base/column.py,sha256=sp3fJstA49FslE2CcgvVFHyi7Jxsxk8qHTd-Z0cAEWc,19932
 sqlframe/base/dataframe.py,sha256=D2N2Kvh_tiF60fYODUikq0xRCJYY4WB2aHbEcq5NIUo,84310
 sqlframe/base/decorators.py,sha256=IhE5xNQDkwJHacCvulq5WpUKyKmXm7dL2A3o5WuKGP4,2131
 sqlframe/base/exceptions.py,sha256=9Uwvqn2eAkDpqm4BrRgbL61qM-GMCbJEMAW8otxO46s,370
 sqlframe/base/function_alternatives.py,sha256=Bs1bwl25fN3Yy9rb4GnUWBGunQ1C_yelkb2yV9DSZIY,53918
-sqlframe/base/functions.py,sha256=iReQ8NW4cwTvgGOXQq6WphC3DQbtKjIHDPJZgWjE614,224862
+sqlframe/base/functions.py,sha256=pL56_rBUwGSYCnNJ9NaD0AT65lOQqNlw1axJxtXrSsA,227187
 sqlframe/base/group.py,sha256=OY4w1WRsCqLgW-Pi7DjF63zbbxSLISCF3qjAbzI2CQ4,4283
 sqlframe/base/normalize.py,sha256=nXAJ5CwxVf4DV0GsH-q1w0p8gmjSMlv96k_ez1eVul8,3880
 sqlframe/base/operations.py,sha256=g-YNcbvNKTOBbYm23GKfB3fmydlR7ZZDAuZUtXIHtzw,4438
@@ -130,8 +130,8 @@ sqlframe/standalone/udf.py,sha256=azmgtUjHNIPs0WMVNId05SHwiYn41MKVBhKXsQJ5dmY,27
 sqlframe/standalone/window.py,sha256=6GKPzuxeSapJakBaKBeT9VpED1ACdjggDv9JRILDyV0,35
 sqlframe/testing/__init__.py,sha256=VVCosQhitU74A3NnE52O4mNtGZONapuEXcc20QmSlnQ,132
 sqlframe/testing/utils.py,sha256=PFsGZpwNUE_4-g_f43_vstTqsK0AQ2lBneb5Eb6NkFo,13008
-sqlframe-3.31.0.dist-info/LICENSE,sha256=VZu79YgW780qxaFJMr0t5ZgbOYEh04xWoxaWOaqIGWk,1068
-sqlframe-3.31.0.dist-info/METADATA,sha256=9D8GGMLw5XslYPUw8U_cEa_Ab4NgB1FRTXW0Wa6lBb0,8987
-sqlframe-3.31.0.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
-sqlframe-3.31.0.dist-info/top_level.txt,sha256=T0_RpoygaZSF6heeWwIDQgaP0varUdSK1pzjeJZRjM8,9
-sqlframe-3.31.0.dist-info/RECORD,,
+sqlframe-3.31.2.dist-info/LICENSE,sha256=VZu79YgW780qxaFJMr0t5ZgbOYEh04xWoxaWOaqIGWk,1068
+sqlframe-3.31.2.dist-info/METADATA,sha256=H7g5gxzMh6Za__8fohB0XWz4xqlgJhWh8fgjCap-kz0,8987
+sqlframe-3.31.2.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
+sqlframe-3.31.2.dist-info/top_level.txt,sha256=T0_RpoygaZSF6heeWwIDQgaP0varUdSK1pzjeJZRjM8,9
+sqlframe-3.31.2.dist-info/RECORD,,