remove callback methods for transformations

Signed-off-by: varun-edachali-dbx <varun.edachali@databricks.com>

Author: varun-edachali-dbx

src/databricks/sql/backend/sea/result_set.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Any, List, Optional, TYPE_CHECKING, Dict
+from typing import Any, List, Optional, TYPE_CHECKING, Dict, Union
 
 import logging
 
@@ -93,8 +93,8 @@ def __init__(
         )
 
         # Initialize metadata columns for post-fetch transformation
-        self._metadata_columns = None
-        self._column_index_mapping = None
+        self._metadata_columns: Optional[List[ResultColumn]] = None
+        self._column_index_mapping: Optional[Dict[int, Union[int, None]]] = None
 
     def _convert_json_types(self, row: List[str]) -> List[Any]:
         """
@@ -301,40 +301,6 @@ def prepare_metadata_columns(self, metadata_columns: List[ResultColumn]) -> None
         self._metadata_columns = metadata_columns
         self._prepare_column_mapping()
 
-    def _populate_columns_from_others(
-        self, result_column: ResultColumn, row_data: Any
-    ) -> Any:
-        """
-        Helper function to populate column data from other columns based on COLUMN_DATA_MAPPING.
-
-        Args:
-            result_column: The result column that needs data
-            row_data: Row data (list for JSON, PyArrow table for Arrow)
-
-        Returns:
-            The value to use for this column, or None if not found
-        """
-        target_column = result_column.column_name
-        if target_column not in COLUMN_DATA_MAPPING:
-            return None
-
-        source_column = COLUMN_DATA_MAPPING[target_column]
-
-        # Find the source column index
-        for idx, col in enumerate(self._metadata_columns):
-            if col.column_name == source_column:
-                source_idx = self._column_index_mapping.get(idx)
-                if source_idx is not None:
-                    # Handle Arrow table format
-                    if hasattr(row_data, "column"):  # PyArrow table
-                        return row_data.column(source_idx).to_pylist()
-                    # Handle JSON row format
-                    else:
-                        return row_data[source_idx]
-                break
-
-        return None
-
     def _prepare_column_mapping(self) -> None:
         """
         Prepare column index mapping for metadata queries.
@@ -353,7 +319,7 @@ def _prepare_column_mapping(self) -> None:
         new_description = []
         self._column_index_mapping = {}  # Maps new index -> old index
 
-        for new_idx, result_column in enumerate(self._metadata_columns):
+        for new_idx, result_column in enumerate(self._metadata_columns or []):
             # Find the corresponding SEA column
             if (
                 result_column.result_set_column_name
@@ -400,8 +366,12 @@ def _transform_arrow_table(self, table: "pyarrow.Table") -> "pyarrow.Table":
         new_columns = []
         column_names = []
 
-        for new_idx, result_column in enumerate(self._metadata_columns):
-            old_idx = self._column_index_mapping.get(new_idx)
+        for new_idx, result_column in enumerate(self._metadata_columns or []):
+            old_idx = (
+                self._column_index_mapping.get(new_idx)
+                if self._column_index_mapping
+                else None
+            )
 
             # Get the source data
             if old_idx is not None:
@@ -410,27 +380,13 @@ def _transform_arrow_table(self, table: "pyarrow.Table") -> "pyarrow.Table":
             else:
                 values = None
 
-            # Special handling for columns that need data from other columns
-            if result_column.result_set_column_name is None:
-                values = self._populate_columns_from_others(result_column, table)
-
             # Apply transformation and create column
             if values is not None:
-                if result_column.transform_value:
-                    transformed_values = [
-                        result_column.transform_value(v) for v in values
-                    ]
-                    column = pyarrow.array(transformed_values)
-                else:
-                    column = pyarrow.array(values)
+                column = pyarrow.array(values)
                 new_columns.append(column)
             else:
                 # Create column with default/transformed values
-                if result_column.transform_value:
-                    default_value = result_column.transform_value(None)
-                    null_array = pyarrow.array([default_value] * table.num_rows)
-                else:
-                    null_array = pyarrow.nulls(table.num_rows)
+                null_array = pyarrow.nulls(table.num_rows)
                 new_columns.append(null_array)
 
             column_names.append(result_column.column_name)
@@ -445,21 +401,17 @@ def _transform_json_rows(self, rows: List[List[str]]) -> List[List[Any]]:
         transformed_rows = []
         for row in rows:
             new_row = []
-            for new_idx, result_column in enumerate(self._metadata_columns):
-                old_idx = self._column_index_mapping.get(new_idx)
+            for new_idx, result_column in enumerate(self._metadata_columns or []):
+                old_idx = (
+                    self._column_index_mapping.get(new_idx)
+                    if self._column_index_mapping
+                    else None
+                )
                 if old_idx is not None:
                     value = row[old_idx]
                 else:
                     value = None
 
-                # Special handling for columns that need data from other columns
-                if result_column.result_set_column_name is None:
-                    value = self._populate_columns_from_others(result_column, row)
-
-                # Apply transformation if defined
-                if result_column.transform_value:
-                    value = result_column.transform_value(value)
-
                 new_row.append(value)
             transformed_rows.append(new_row)
         return transformed_rows

src/databricks/sql/backend/sea/utils/metadata_mappings.py

@@ -1,141 +1,79 @@
 from databricks.sql.backend.sea.utils.result_column import ResultColumn
-from databricks.sql.backend.sea.utils.metadata_transformations import (
-    transform_table_type,
-    transform_is_nullable,
-    transform_nullable_to_int,
-    transform_remarks_default,
-    transform_numeric_default_zero,
-    transform_ordinal_position_offset,
-    calculate_data_type,
-    calculate_buffer_length,
-    always_null,
-    always_null_int,
-    always_null_smallint,
-    identity,
-)
 
 
 class MetadataColumnMappings:
     """Column mappings for metadata queries following JDBC specification."""
 
     # Common columns used across multiple metadata queries
     # FIX 1: Catalog columns - swap the mappings
-    CATALOG_COLUMN = ResultColumn(
-        "TABLE_CAT", "catalogName", "string", transform_value=identity
-    )
-    CATALOG_COLUMN_FOR_GET_CATALOGS = ResultColumn(
-        "TABLE_CAT", "catalog", "string", transform_value=identity
-    )
+    CATALOG_COLUMN = ResultColumn("TABLE_CAT", "catalogName", "string")
+    CATALOG_COLUMN_FOR_GET_CATALOGS = ResultColumn("TABLE_CAT", "catalog", "string")
     # Remove CATALOG_COLUMN_FOR_TABLES - will use CATALOG_COLUMN instead
 
-    SCHEMA_COLUMN = ResultColumn(
-        "TABLE_SCHEM", "namespace", "string", transform_value=identity
-    )
-    SCHEMA_COLUMN_FOR_GET_SCHEMA = ResultColumn(
-        "TABLE_SCHEM", "databaseName", "string", transform_value=identity
-    )
-    TABLE_NAME_COLUMN = ResultColumn(
-        "TABLE_NAME", "tableName", "string", transform_value=identity
-    )
-    TABLE_TYPE_COLUMN = ResultColumn(
-        "TABLE_TYPE", "tableType", "string", transform_value=transform_table_type
-    )
-    REMARKS_COLUMN = ResultColumn(
-        "REMARKS", "remarks", "string", transform_value=transform_remarks_default
-    )
+    SCHEMA_COLUMN = ResultColumn("TABLE_SCHEM", "namespace", "string")
+    SCHEMA_COLUMN_FOR_GET_SCHEMA = ResultColumn("TABLE_SCHEM", "databaseName", "string")
+    TABLE_NAME_COLUMN = ResultColumn("TABLE_NAME", "tableName", "string")
+    TABLE_TYPE_COLUMN = ResultColumn("TABLE_TYPE", "tableType", "string")
+    REMARKS_COLUMN = ResultColumn("REMARKS", "remarks", "string")
 
     # Columns specific to getColumns()
-    COLUMN_NAME_COLUMN = ResultColumn(
-        "COLUMN_NAME", "col_name", "string", transform_value=identity
-    )
+    COLUMN_NAME_COLUMN = ResultColumn("COLUMN_NAME", "col_name", "string")
     DATA_TYPE_COLUMN = ResultColumn(
-        "DATA_TYPE", None, "int", transform_value=calculate_data_type
+        "DATA_TYPE", None, "int"
     )  # Calculated from columnType
-    TYPE_NAME_COLUMN = ResultColumn(
-        "TYPE_NAME", "columnType", "string", transform_value=identity
-    )
+    TYPE_NAME_COLUMN = ResultColumn("TYPE_NAME", "columnType", "string")
 
     # FIX 5: SEA actually provides these columns
-    COLUMN_SIZE_COLUMN = ResultColumn(
-        "COLUMN_SIZE", "columnSize", "int", transform_value=identity
-    )
+    COLUMN_SIZE_COLUMN = ResultColumn("COLUMN_SIZE", "columnSize", "int")
     DECIMAL_DIGITS_COLUMN = ResultColumn(
         "DECIMAL_DIGITS",
         "decimalDigits",
         "int",
-        transform_value=transform_numeric_default_zero,
-    )
-    NUM_PREC_RADIX_COLUMN = ResultColumn(
-        "NUM_PREC_RADIX", "radix", "int", transform_value=transform_numeric_default_zero
     )
+    NUM_PREC_RADIX_COLUMN = ResultColumn("NUM_PREC_RADIX", "radix", "int")
     ORDINAL_POSITION_COLUMN = ResultColumn(
         "ORDINAL_POSITION",
         "ordinalPosition",
         "int",
-        transform_value=transform_ordinal_position_offset,
     )
 
     NULLABLE_COLUMN = ResultColumn(
-        "NULLABLE", None, "int", transform_value=transform_nullable_to_int
+        "NULLABLE", None, "int"
     )  # Calculated from isNullable
     COLUMN_DEF_COLUMN = ResultColumn(
-        "COLUMN_DEF", "columnType", "string", transform_value=identity
+        "COLUMN_DEF", "columnType", "string"
     )  # Note: duplicate mapping
-    SQL_DATA_TYPE_COLUMN = ResultColumn(
-        "SQL_DATA_TYPE", None, "int", transform_value=always_null_int
-    )
-    SQL_DATETIME_SUB_COLUMN = ResultColumn(
-        "SQL_DATETIME_SUB", None, "int", transform_value=always_null_int
-    )
-    CHAR_OCTET_LENGTH_COLUMN = ResultColumn(
-        "CHAR_OCTET_LENGTH", None, "int", transform_value=always_null_int
-    )
-    IS_NULLABLE_COLUMN = ResultColumn(
-        "IS_NULLABLE", "isNullable", "string", transform_value=transform_is_nullable
-    )
+    SQL_DATA_TYPE_COLUMN = ResultColumn("SQL_DATA_TYPE", None, "int")
+    SQL_DATETIME_SUB_COLUMN = ResultColumn("SQL_DATETIME_SUB", None, "int")
+    CHAR_OCTET_LENGTH_COLUMN = ResultColumn("CHAR_OCTET_LENGTH", None, "int")
+    IS_NULLABLE_COLUMN = ResultColumn("IS_NULLABLE", "isNullable", "string")
 
     # Columns for getTables() that don't exist in SEA
-    TYPE_CAT_COLUMN = ResultColumn(
-        "TYPE_CAT", None, "string", transform_value=always_null
-    )
-    TYPE_SCHEM_COLUMN = ResultColumn(
-        "TYPE_SCHEM", None, "string", transform_value=always_null
-    )
-    TYPE_NAME_COLUMN = ResultColumn(
-        "TYPE_NAME", None, "string", transform_value=always_null
-    )
+    TYPE_CAT_COLUMN = ResultColumn("TYPE_CAT", None, "string")
+    TYPE_SCHEM_COLUMN = ResultColumn("TYPE_SCHEM", None, "string")
+    TYPE_NAME_COLUMN = ResultColumn("TYPE_NAME", None, "string")
     SELF_REFERENCING_COL_NAME_COLUMN = ResultColumn(
-        "SELF_REFERENCING_COL_NAME", None, "string", transform_value=always_null
-    )
-    REF_GENERATION_COLUMN = ResultColumn(
-        "REF_GENERATION", None, "string", transform_value=always_null
+        "SELF_REFERENCING_COL_NAME", None, "string"
     )
+    REF_GENERATION_COLUMN = ResultColumn("REF_GENERATION", None, "string")
 
     # FIX 8: Scope columns (always null per JDBC)
-    SCOPE_CATALOG_COLUMN = ResultColumn(
-        "SCOPE_CATALOG", None, "string", transform_value=always_null
-    )
-    SCOPE_SCHEMA_COLUMN = ResultColumn(
-        "SCOPE_SCHEMA", None, "string", transform_value=always_null
-    )
-    SCOPE_TABLE_COLUMN = ResultColumn(
-        "SCOPE_TABLE", None, "string", transform_value=always_null
-    )
-    SOURCE_DATA_TYPE_COLUMN = ResultColumn(
-        "SOURCE_DATA_TYPE", None, "smallint", transform_value=always_null_smallint
-    )
+    SCOPE_CATALOG_COLUMN = ResultColumn("SCOPE_CATALOG", None, "string")
+    SCOPE_SCHEMA_COLUMN = ResultColumn("SCOPE_SCHEMA", None, "string")
+    SCOPE_TABLE_COLUMN = ResultColumn("SCOPE_TABLE", None, "string")
+    SOURCE_DATA_TYPE_COLUMN = ResultColumn("SOURCE_DATA_TYPE", None, "smallint")
 
     # FIX 9 & 10: Auto increment and generated columns
     IS_AUTO_INCREMENT_COLUMN = ResultColumn(
-        "IS_AUTOINCREMENT", "isAutoIncrement", "string", transform_value=identity
+        "IS_AUTOINCREMENT", "isAutoIncrement", "string"
     )  # No underscore!
     IS_GENERATED_COLUMN = ResultColumn(
-        "IS_GENERATEDCOLUMN", "isGenerated", "string", transform_value=identity
+        "IS_GENERATEDCOLUMN", "isGenerated", "string"
     )  # SEA provides this
 
     # FIX 11: Buffer length column
     BUFFER_LENGTH_COLUMN = ResultColumn(
-        "BUFFER_LENGTH", None, "int", transform_value=always_null_int
+        "BUFFER_LENGTH", None, "int"
     )  # Always null per JDBC
 
     # Column lists for each metadata operation
@@ -144,7 +82,7 @@ class MetadataColumnMappings:
     SCHEMA_COLUMNS = [
         SCHEMA_COLUMN_FOR_GET_SCHEMA,
         ResultColumn(
-            "TABLE_CATALOG", None, "string", transform_value=always_null
+            "TABLE_CATALOG", None, "string"
         ),  # Will need special population logic
     ]