Removing input name remover shuttle (#426)

This shuttle is now defunct as other functionalities render it obsolete.

Author: knassre-bodo

pydough/conversion/agg_split.py

@@ -13,7 +13,6 @@
     CallExpression,
     ColumnReference,
     ColumnReferenceFinder,
-    ColumnReferenceInputNameRemover,
     Join,
     JoinType,
     LiteralExpression,
@@ -23,6 +22,7 @@
 )
 from pydough.relational.rel_util import (
     ExpressionTranspositionShuttle,
+    add_input_name,
     extract_equijoin_keys,
     fetch_or_insert,
 )
@@ -173,7 +173,6 @@ def transpose_aggregate_join(
     need_projection: bool = False
 
     finder: ColumnReferenceFinder = ColumnReferenceFinder()
-    alias_remover: ColumnReferenceInputNameRemover = ColumnReferenceInputNameRemover()
     transposer: ExpressionTranspositionShuttle = ExpressionTranspositionShuttle(
         join, False
     )
@@ -245,8 +244,8 @@ def transpose_aggregate_join(
             agg_input_name
         }:
             if isinstance(transposed_agg_key, ColumnReference):
-                input_keys[transposed_agg_key.name] = transposed_agg_key.accept_shuttle(
-                    alias_remover
+                input_keys[transposed_agg_key.name] = add_input_name(
+                    transposed_agg_key, None
                 )
             else:
                 if agg_key_name in join.columns and (
@@ -255,9 +254,7 @@ def transpose_aggregate_join(
                     # An edge cases that is theoretically possible but never
                     # encountered so far, and where the behavior is undefined.
                     raise NotImplementedError("Undefined behavior")
-                input_keys[agg_key_name] = transposed_agg_key.accept_shuttle(
-                    alias_remover
-                )
+                input_keys[agg_key_name] = add_input_name(transposed_agg_key, None)
                 join.columns[agg_key_name] = ColumnReference(
                     agg_key_name, agg_key.data_type, agg_input_name
                 )

pydough/relational/__init__.py

@@ -5,7 +5,6 @@
     "ColumnReference",
     "ColumnReferenceFinder",
     "ColumnReferenceInputNameModifier",
-    "ColumnReferenceInputNameRemover",
     "CorrelatedReference",
     "EmptySingleton",
     "ExpressionSortInfo",
@@ -34,7 +33,6 @@
     ColumnReference,
     ColumnReferenceFinder,
     ColumnReferenceInputNameModifier,
-    ColumnReferenceInputNameRemover,
     CorrelatedReference,
     ExpressionSortInfo,
     LiteralExpression,

pydough/relational/relational_expressions/README.md

@@ -57,10 +57,6 @@ The relational_expressions module provides functionality to define and manage va
 
 - `RelationalExpressionShuttle`: Specialized form of the visitor pattern that returns a relational expression. This is used to handle the common case where we need to modify a type of input.
 
-### [column_reference_input_name_remover.py](column_reference_input_name_remover.py)
-
-- `ColumnReferenceInputNameRemover`: Shuttle implementation designed to remove the input name from any column reference whose name is not found in the given set.
-
 ### [column_reference_input_name_modifier.py](column_reference_input_name_modifier.py)
 
 - `ColumnReferenceInputNameModifier`: Shuttle implementation designed to update all uses of a column reference's input name to a new input name based on a dictionary.

pydough/relational/relational_expressions/__init__.py

@@ -8,7 +8,6 @@
     "ColumnReference",
     "ColumnReferenceFinder",
     "ColumnReferenceInputNameModifier",
-    "ColumnReferenceInputNameRemover",
     "CorrelatedReference",
     "CorrelatedReferenceFinder",
     "ExpressionSortInfo",
@@ -23,7 +22,6 @@
 from .column_reference import ColumnReference
 from .column_reference_finder import ColumnReferenceFinder
 from .column_reference_input_name_modifier import ColumnReferenceInputNameModifier
-from .column_reference_input_name_remover import ColumnReferenceInputNameRemover
 from .correlated_reference import CorrelatedReference
 from .correlated_reference_finder import CorrelatedReferenceFinder
 from .expression_sort_info import ExpressionSortInfo

pydough/relational/relational_expressions/column_reference_input_name_remover.py

@@ -4,7 +4,6 @@
 """
 
 import warnings
-from collections import defaultdict
 
 from sqlglot.expressions import AggFunc as SQLGlotAggFunc
 from sqlglot.expressions import Alias as SQLGlotAlias
@@ -23,13 +22,11 @@
     CallExpression,
     ColumnReference,
     ColumnReferenceInputNameModifier,
-    ColumnReferenceInputNameRemover,
     CorrelatedReference,
     EmptySingleton,
     ExpressionSortInfo,
     Filter,
     Join,
-    JoinType,
     Limit,
     LiteralExpression,
     Project,
@@ -92,9 +89,6 @@ def __init__(
         self._alias_modifier: ColumnReferenceInputNameModifier = (
             ColumnReferenceInputNameModifier()
         )
-        self._alias_remover: ColumnReferenceInputNameRemover = (
-            ColumnReferenceInputNameRemover()
-        )
         # Counter for generating unique table alias.
         self._alias_counter: int = 0
 
@@ -399,29 +393,13 @@ def visit_join(self, join: Join) -> None:
         self.visit_inputs(join)
         inputs: list[Select] = [self._stack.pop() for _ in range(len(join.inputs))]
         inputs.reverse()
-        # Compute a dictionary to find all duplicate names.
-        seen_names: dict[str, int] = defaultdict(int)
-        for idx, input in enumerate(join.inputs):
-            occurrences: int = 1
-            if join.join_type in (JoinType.SEMI, JoinType.ANTI):
-                # For SEMI and ANTI joins, we must keep the columns input names
-                # from the left side and treat them as if they appear in the
-                # RHS.
-                occurrences = 2
-            for column in input.columns.keys():
-                seen_names[column] += occurrences
-        # Only keep duplicate names.
-        kept_names = {key for key, value in seen_names.items() if value > 1}
         for i in range(len(join.inputs)):
             input_name = join.default_input_aliases[i]
             if input_name not in alias_map:
                 alias_map[input_name] = self._generate_table_alias()
-        self._alias_remover.set_kept_names(kept_names)
         self._alias_modifier.set_map(alias_map)
         columns = {
-            alias: col.accept_shuttle(self._alias_remover).accept_shuttle(
-                self._alias_modifier
-            )
+            alias: col.accept_shuttle(self._alias_modifier)
             for alias, col in join.columns.items()
         }
         column_exprs = [

pydough/sqlglot/sqlglot_relational_visitor.py

@@ -26,8 +26,8 @@ WITH _t5 AS (
     3
 ), _s11 AS (
   SELECT
-    _t6.l_orderkey,
     _t6.l_linenumber,
+    _t6.l_orderkey,
     orders.o_orderkey
   FROM _t5 AS _t6
   JOIN tpch.orders AS orders