window() has been deprecated since DF50

Author: timsaucer

crates/core/src/functions.rs

@@ -18,20 +18,14 @@
 use std::collections::HashMap;
 
 use datafusion::common::{Column, ScalarValue, TableReference};
-use datafusion::execution::FunctionRegistry;
-use datafusion::functions_aggregate::all_default_aggregate_functions;
-use datafusion::functions_window::all_default_window_functions;
-use datafusion::logical_expr::expr::{
-    Alias, FieldMetadata, NullTreatment as DFNullTreatment, WindowFunction, WindowFunctionParams,
-};
-use datafusion::logical_expr::{Expr, ExprFunctionExt, WindowFrame, WindowFunctionDefinition, lit};
+use datafusion::logical_expr::expr::{Alias, FieldMetadata, NullTreatment as DFNullTreatment};
+use datafusion::logical_expr::{Expr, ExprFunctionExt, lit};
 use datafusion::{functions, functions_aggregate, functions_window};
 use pyo3::prelude::*;
 use pyo3::wrap_pyfunction;
 
 use crate::common::data_type::{NullTreatment, PyScalarValue};
-use crate::context::PySessionContext;
-use crate::errors::{PyDataFusionError, PyDataFusionResult};
+use crate::errors::PyDataFusionResult;
 use crate::expr::PyExpr;
 use crate::expr::conditional_expr::PyCaseBuilder;
 use crate::expr::sort_expr::{PySortExpr, to_sort_expressions};
@@ -306,126 +300,6 @@ fn when(when: PyExpr, then: PyExpr) -> PyResult<PyCaseBuilder> {
     Ok(PyCaseBuilder::new(None).when(when, then))
 }
 
-/// Helper function to find the appropriate window function.
-///
-/// Search procedure:
-/// 1) Search built in window functions, which are being deprecated.
-/// 1) If a session context is provided:
-///      1) search User Defined Aggregate Functions (UDAFs)
-///      1) search registered window functions
-///      1) search registered aggregate functions
-/// 1) If no function has been found, search default aggregate functions.
-///
-/// NOTE: we search the built-ins first because the `UDAF` versions currently do not have the same behavior.
-fn find_window_fn(
-    name: &str,
-    ctx: Option<PySessionContext>,
-) -> PyDataFusionResult<WindowFunctionDefinition> {
-    if let Some(ctx) = ctx {
-        // search UDAFs
-        let udaf = ctx
-            .ctx
-            .udaf(name)
-            .map(WindowFunctionDefinition::AggregateUDF)
-            .ok();
-
-        if let Some(udaf) = udaf {
-            return Ok(udaf);
-        }
-
-        let session_state = ctx.ctx.state();
-
-        // search registered window functions
-        let window_fn = session_state
-            .window_functions()
-            .get(name)
-            .map(|f| WindowFunctionDefinition::WindowUDF(f.clone()));
-
-        if let Some(window_fn) = window_fn {
-            return Ok(window_fn);
-        }
-
-        // search registered aggregate functions
-        let agg_fn = session_state
-            .aggregate_functions()
-            .get(name)
-            .map(|f| WindowFunctionDefinition::AggregateUDF(f.clone()));
-
-        if let Some(agg_fn) = agg_fn {
-            return Ok(agg_fn);
-        }
-    }
-
-    // search default aggregate functions
-    let agg_fn = all_default_aggregate_functions()
-        .iter()
-        .find(|v| v.name() == name || v.aliases().contains(&name.to_string()))
-        .map(|f| WindowFunctionDefinition::AggregateUDF(f.clone()));
-
-    if let Some(agg_fn) = agg_fn {
-        return Ok(agg_fn);
-    }
-
-    // search default window functions
-    let window_fn = all_default_window_functions()
-        .iter()
-        .find(|v| v.name() == name || v.aliases().contains(&name.to_string()))
-        .map(|f| WindowFunctionDefinition::WindowUDF(f.clone()));
-
-    if let Some(window_fn) = window_fn {
-        return Ok(window_fn);
-    }
-
-    Err(PyDataFusionError::Common(format!(
-        "window function `{name}` not found"
-    )))
-}
-
-/// Creates a new Window function expression
-#[allow(clippy::too_many_arguments)]
-#[pyfunction]
-#[pyo3(signature = (name, args, partition_by=None, order_by=None, window_frame=None, filter=None, distinct=false, ctx=None))]
-fn window(
-    name: &str,
-    args: Vec<PyExpr>,
-    partition_by: Option<Vec<PyExpr>>,
-    order_by: Option<Vec<PySortExpr>>,
-    window_frame: Option<PyWindowFrame>,
-    filter: Option<PyExpr>,
-    distinct: bool,
-    ctx: Option<PySessionContext>,
-) -> PyResult<PyExpr> {
-    let fun = find_window_fn(name, ctx)?;
-
-    let window_frame = window_frame
-        .map(|w| w.into())
-        .unwrap_or(WindowFrame::new(order_by.as_ref().map(|v| !v.is_empty())));
-    let filter = filter.map(|f| f.expr.into());
-
-    Ok(PyExpr {
-        expr: datafusion::logical_expr::Expr::WindowFunction(Box::new(WindowFunction {
-            fun,
-            params: WindowFunctionParams {
-                args: args.into_iter().map(|x| x.expr).collect::<Vec<_>>(),
-                partition_by: partition_by
-                    .unwrap_or_default()
-                    .into_iter()
-                    .map(|x| x.expr)
-                    .collect::<Vec<_>>(),
-                order_by: order_by
-                    .unwrap_or_default()
-                    .into_iter()
-                    .map(|x| x.into())
-                    .collect::<Vec<_>>(),
-                window_frame,
-                filter,
-                distinct,
-                null_treatment: None,
-            },
-        })),
-    })
-}
-
 // Generates a [pyo3] wrapper for associated aggregate functions.
 // All of the builder options are exposed to the python internal
 // function and we rely on the wrappers to only use those that
@@ -1186,7 +1060,6 @@ pub(crate) fn init_module(m: &Bound<'_, PyModule>) -> PyResult<()> {
     m.add_wrapped(wrap_pyfunction!(self::uuid))?; // Use self to avoid name collision
     m.add_wrapped(wrap_pyfunction!(var_pop))?;
     m.add_wrapped(wrap_pyfunction!(var_sample))?;
-    m.add_wrapped(wrap_pyfunction!(window))?;
     m.add_wrapped(wrap_pyfunction!(regr_avgx))?;
     m.add_wrapped(wrap_pyfunction!(regr_avgy))?;
     m.add_wrapped(wrap_pyfunction!(regr_count))?;

python/datafusion/functions.py

@@ -18,7 +18,7 @@
 
 from __future__ import annotations
 
-from typing import TYPE_CHECKING, Any
+from typing import Any
 
 import pyarrow as pa
 
@@ -29,19 +29,11 @@
     Expr,
     SortExpr,
     SortKey,
-    WindowFrame,
     expr_list_to_raw_expr_list,
     sort_list_to_raw_sort_list,
     sort_or_default,
 )
 
-try:
-    from warnings import deprecated  # Python 3.13+
-except ImportError:
-    from typing_extensions import deprecated  # Python 3.12
-
-if TYPE_CHECKING:
-    from datafusion.context import SessionContext
 __all__ = [
     "abs",
     "acos",
@@ -339,8 +331,6 @@
     "var_sample",
     "version",
     "when",
-    # Window Functions
-    "window",
 ]
 
 
@@ -664,49 +654,6 @@ def when(when: Expr, then: Expr) -> CaseBuilder:
     return CaseBuilder(f.when(when.expr, then.expr))
 
 
-@deprecated("Prefer to call Expr.over() instead")
-def window(
-    name: str,
-    args: list[Expr],
-    partition_by: list[Expr] | Expr | None = None,
-    order_by: list[SortKey] | SortKey | None = None,
-    window_frame: WindowFrame | None = None,
-    filter: Expr | None = None,
-    distinct: bool = False,
-    ctx: SessionContext | None = None,
-) -> Expr:
-    """Creates a new Window function expression.
-
-    This interface will soon be deprecated. Instead of using this interface,
-    users should call the window functions directly. For example, to perform a
-    lag use::
-
-        df.select(functions.lag(col("a")).partition_by(col("b")).build())
-
-    The ``order_by`` parameter accepts column names or expressions, e.g.::
-
-        window("lag", [col("a")], order_by="ts")
-    """
-    args = [a.expr for a in args]
-    partition_by_raw = expr_list_to_raw_expr_list(partition_by)
-    order_by_raw = sort_list_to_raw_sort_list(order_by)
-    window_frame = window_frame.window_frame if window_frame is not None else None
-    ctx = ctx.ctx if ctx is not None else None
-    filter_raw = filter.expr if filter is not None else None
-    return Expr(
-        f.window(
-            name,
-            args,
-            partition_by=partition_by_raw,
-            order_by=order_by_raw,
-            window_frame=window_frame,
-            ctx=ctx,
-            filter=filter_raw,
-            distinct=distinct,
-        )
-    )
-
-
 # scalar functions
 def abs(arg: Expr) -> Expr:
     """Return the absolute value of a given number.