pylegend 0.2.2__py3-none-any.whl → 0.4.0__py3-none-any.whl

pylegend/core/tds/legendql_api/frames/functions/legendql_api_function_helpers.py

@@ -0,0 +1,137 @@
+# Copyright 2025 Goldman Sachs
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from pylegend._typing import (
+    PyLegendList,
+    PyLegendSequence,
+    PyLegendCallable,
+    PyLegendUnion,
+)
+from pylegend.core.language import PyLegendColumnExpression
+from pylegend.core.language.legendql_api.legendql_api_custom_expressions import (
+    LegendQLApiPrimitive,
+    LegendQLApiSortInfo,
+    LegendQLApiSortDirection,
+)
+from pylegend.core.language.legendql_api.legendql_api_tds_row import LegendQLApiTdsRow
+from pylegend.core.tds.legendql_api.frames.legendql_api_base_tds_frame import LegendQLApiBaseTdsFrame
+
+
+__all__: PyLegendSequence[str] = [
+    "infer_columns_from_frame",
+    "infer_sorts_from_frame",
+]
+
+
+def infer_columns_from_frame(
+    base_frame: LegendQLApiBaseTdsFrame,
+    columns: PyLegendUnion[
+        str,
+        PyLegendList[str],
+        PyLegendCallable[
+            [LegendQLApiTdsRow],
+            PyLegendUnion[LegendQLApiPrimitive, PyLegendList[LegendQLApiPrimitive]]
+        ]
+    ],
+    context: str
+) -> PyLegendList[str]:
+    if isinstance(columns, str):
+        return [columns]
+    if isinstance(columns, list) and all([isinstance(c, str) for c in columns]):
+        return columns
+
+    if isinstance(columns, type(lambda x: 0)) and (columns.__code__.co_argcount == 1):
+        tds_row = LegendQLApiTdsRow.from_tds_frame("frame", base_frame)
+        try:
+            result = columns(tds_row)
+        except Exception as e:
+            raise RuntimeError(
+                f"{context} argument lambda incompatible. Error occurred while evaluating. Message: " + str(e)
+            ) from e
+
+        list_result = result if isinstance(result, list) else [result]
+
+        columns_list = []
+        for (i, r) in enumerate(list_result):
+            if isinstance(r, LegendQLApiPrimitive) and isinstance(r.value(), PyLegendColumnExpression):
+                col_expr: PyLegendColumnExpression = r.value()
+                columns_list.append(col_expr.get_column())
+            else:
+                raise TypeError(
+                    f"{context} argument lambda incompatible. Columns can be simple column expressions "
+                    f"(E.g - lambda r: [r.column1, r.column2, r['column with spaces']). "
+                    f"Element at index {i} (0-indexed) is incompatible"
+                )
+        return columns_list
+
+    raise TypeError(f"{context} argument can either be a list of strings (column names) or "
+                    f"a lambda function which takes one argument (LegendQLApiTdsRow)")
+
+
+def infer_sorts_from_frame(
+    base_frame: LegendQLApiBaseTdsFrame,
+    sort_infos: PyLegendUnion[
+        str,
+        PyLegendList[str],
+        PyLegendCallable[
+            [LegendQLApiTdsRow],
+            PyLegendUnion[
+                LegendQLApiPrimitive,
+                LegendQLApiSortInfo,
+                PyLegendList[PyLegendUnion[LegendQLApiPrimitive, LegendQLApiSortInfo]],
+            ]
+        ]
+    ],
+    context: str
+) -> PyLegendList[LegendQLApiSortInfo]:
+    tds_row = LegendQLApiTdsRow.from_tds_frame("frame", base_frame)
+    sort_info_list = []
+
+    if isinstance(sort_infos, str):
+        col_expr1: PyLegendColumnExpression = tds_row[sort_infos].value()  # type: ignore
+        sort_info_list.append(LegendQLApiSortInfo(col_expr1, LegendQLApiSortDirection.ASC))
+
+    elif isinstance(sort_infos, list) and all([isinstance(s, str) for s in sort_infos]):
+        for s in sort_infos:
+            col_expr2: PyLegendColumnExpression = tds_row[s].value()  # type: ignore
+            sort_info_list.append(LegendQLApiSortInfo(col_expr2, LegendQLApiSortDirection.ASC))
+
+    elif isinstance(sort_infos, type(lambda x: 0)) and (sort_infos.__code__.co_argcount == 1):
+        try:
+            result = sort_infos(tds_row)
+        except Exception as e:
+            raise RuntimeError(
+                f"{context} argument lambda incompatible. "
+                "Error occurred while evaluating. Message: " + str(e)
+            ) from e
+
+        list_result = result if isinstance(result, list) else [result]
+        for (i, r) in enumerate(list_result):
+            if isinstance(r, LegendQLApiPrimitive) and isinstance(r.value(), PyLegendColumnExpression):
+                col_expr3: PyLegendColumnExpression = r.value()
+                sort_info_list.append(LegendQLApiSortInfo(col_expr3, LegendQLApiSortDirection.ASC))
+            elif isinstance(r, LegendQLApiSortInfo):
+                sort_info_list.append(r)
+            else:
+                raise TypeError(
+                    f"{context} argument lambda incompatible. Columns can be simple column "
+                    f"expressions or sort infos. "
+                    f"(E.g - lambda r: [r.column1, r['column with spaces'].descending(), r.column3.ascending()). "
+                    f"Element at index {i} (0-indexed) is incompatible"
+                )
+    else:
+        raise TypeError(f"{context} argument can either be a list of strings (column names) or "
+                        "a lambda function which takes one argument (LegendQLApiTdsRow)")
+
+    return sort_info_list

pylegend/core/tds/legendql_api/frames/functions/legendql_api_groupby_function.py

@@ -0,0 +1,256 @@
+# Copyright 2025 Goldman Sachs
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from datetime import date, datetime
+from pylegend._typing import (
+    PyLegendList,
+    PyLegendSequence,
+    PyLegendTuple,
+    PyLegendCallable,
+    PyLegendUnion,
+)
+from pylegend.core.language import (
+    PyLegendPrimitiveOrPythonPrimitive,
+    PyLegendPrimitiveCollection,
+    PyLegendPrimitive,
+    create_primitive_collection,
+    convert_literal_to_literal_expression
+)
+from pylegend.core.language.legendql_api.legendql_api_custom_expressions import LegendQLApiPrimitive
+from pylegend.core.language.legendql_api.legendql_api_tds_row import LegendQLApiTdsRow
+from pylegend.core.tds.abstract.function_helpers import tds_column_for_primitive
+from pylegend.core.tds.legendql_api.frames.functions.legendql_api_function_helpers import infer_columns_from_frame
+from pylegend.core.tds.legendql_api.frames.legendql_api_applied_function_tds_frame import LegendQLApiAppliedFunction
+from pylegend.core.tds.sql_query_helpers import copy_query, create_sub_query
+from pylegend.core.sql.metamodel import (
+    QuerySpecification,
+    SingleColumn,
+    SelectItem
+)
+from pylegend.core.tds.tds_column import TdsColumn
+from pylegend.core.tds.tds_frame import FrameToSqlConfig
+from pylegend.core.tds.tds_frame import FrameToPureConfig
+from pylegend.core.tds.legendql_api.frames.legendql_api_base_tds_frame import LegendQLApiBaseTdsFrame
+from pylegend.core.language.shared.helpers import escape_column_name, generate_pure_lambda
+
+__all__: PyLegendSequence[str] = [
+    "LegendQLApiGroupByFunction"
+]
+
+
+class LegendQLApiGroupByFunction(LegendQLApiAppliedFunction):
+    __base_frame: LegendQLApiBaseTdsFrame
+    __grouping_column_name_list: PyLegendList[str]
+    __aggregates_list: PyLegendList[PyLegendTuple[str, PyLegendPrimitiveOrPythonPrimitive, PyLegendPrimitive]]
+
+    @classmethod
+    def name(cls) -> str:
+        return "groupBy"
+
+    def __init__(
+            self,
+            base_frame: LegendQLApiBaseTdsFrame,
+            grouping_columns: PyLegendUnion[
+                str,
+                PyLegendList[str],
+                PyLegendCallable[
+                    [LegendQLApiTdsRow],
+                    PyLegendUnion[LegendQLApiPrimitive, PyLegendList[LegendQLApiPrimitive]]
+                ]
+            ],
+            aggregate_specifications: PyLegendUnion[
+                PyLegendTuple[
+                    str,
+                    PyLegendCallable[[LegendQLApiTdsRow], PyLegendPrimitiveOrPythonPrimitive],
+                    PyLegendCallable[[PyLegendPrimitiveCollection], PyLegendPrimitive]
+                ],
+                PyLegendList[
+                    PyLegendTuple[
+                        str,
+                        PyLegendCallable[[LegendQLApiTdsRow], PyLegendPrimitiveOrPythonPrimitive],
+                        PyLegendCallable[[PyLegendPrimitiveCollection], PyLegendPrimitive]
+                    ]
+                ]
+            ]
+    ) -> None:
+        self.__base_frame = base_frame
+        self.__grouping_column_name_list = infer_columns_from_frame(
+            base_frame, grouping_columns, "'group_by' function grouping columns"
+        )
+
+        tds_row = LegendQLApiTdsRow.from_tds_frame("r", self.__base_frame)
+        list_result = (
+            aggregate_specifications if isinstance(aggregate_specifications, list) else [aggregate_specifications]
+        )
+        aggregates_list: PyLegendList[PyLegendTuple[str, PyLegendPrimitiveOrPythonPrimitive, PyLegendPrimitive]] = []
+        for (i, agg_spec) in enumerate(list_result):
+            error = (
+                "'group_by' function aggregate specifications incompatible. "
+                "Each aggregate specification should be a triplet with first element being the aggregation column "
+                "name, second element being a mapper function (single argument lambda) and third element being the "
+                "aggregation function (single argument lambda). "
+                "E.g - ('count_col', lambda r: r['col1'], lambda c: c.count()). "
+                f"Element at index {i} (0-indexed) is incompatible"
+            )
+
+            if isinstance(agg_spec, tuple) and isinstance(agg_spec[0], str):
+
+                if not isinstance(agg_spec[1], type(lambda x: 0)) or (agg_spec[1].__code__.co_argcount != 1):
+                    raise TypeError(error)
+
+                try:
+                    map_result = agg_spec[1](tds_row)
+                except Exception as e:
+                    raise RuntimeError(
+                        "'group_by' function aggregate specifications incompatible. "
+                        f"Error occurred while evaluating mapper lambda in the aggregate specification at "
+                        f"index {i} (0-indexed). Message: " + str(e)
+                    ) from e
+
+                if not isinstance(map_result, (int, float, bool, str, date, datetime, PyLegendPrimitive)):
+                    raise TypeError(
+                        "'group_by' function aggregate specifications incompatible. "
+                        f"Mapper lambda in the aggregate specification at index {i} (0-indexed) "
+                        f"returns non-primitive - {str(type(map_result))}"
+                    )
+
+                collection = create_primitive_collection(map_result)
+
+                if not isinstance(agg_spec[2], type(lambda x: 0)) or (agg_spec[2].__code__.co_argcount != 1):
+                    raise TypeError(error)
+
+                try:
+                    agg_result = agg_spec[2](collection)
+                except Exception as e:
+                    raise RuntimeError(
+                        "'group_by' function aggregate specifications incompatible. "
+                        f"Error occurred while evaluating aggregation lambda in the aggregate specification at "
+                        f"index {i} (0-indexed). Message: " + str(e)
+                    ) from e
+
+                if not isinstance(agg_result, PyLegendPrimitive):
+                    raise TypeError(
+                        "'group_by' function aggregate specifications incompatible. "
+                        f"Aggregation lambda in the aggregate specification at index {i} (0-indexed) "
+                        f"returns non-primitive - {str(type(agg_result))}"
+                    )
+
+                aggregates_list.append((agg_spec[0], map_result, agg_result))
+
+            else:
+                raise TypeError(error)
+
+        self.__aggregates_list = aggregates_list
+
+    def to_sql(self, config: FrameToSqlConfig) -> QuerySpecification:
+        db_extension = config.sql_to_string_generator().get_db_extension()
+        base_query = self.__base_frame.to_sql_query_object(config)
+
+        should_create_sub_query = (len(base_query.groupBy) > 0) or base_query.select.distinct or \
+                                  (base_query.offset is not None) or (base_query.limit is not None)
+
+        columns_to_retain = [db_extension.quote_identifier(x) for x in self.__grouping_column_name_list]
+        if should_create_sub_query:
+            new_query = create_sub_query(base_query, config, "root")
+        else:
+            new_query = copy_query(base_query)
+
+        new_cols_with_index: PyLegendList[PyLegendTuple[int, 'SelectItem']] = []
+        for col in new_query.select.selectItems:
+            if not isinstance(col, SingleColumn):
+                raise ValueError("Group By operation not supported for queries "
+                                 "with columns other than SingleColumn")  # pragma: no cover
+            if col.alias is None:
+                raise ValueError("Group By operation not supported for queries "
+                                 "with SingleColumns with missing alias")  # pragma: no cover
+            if col.alias in columns_to_retain:
+                new_cols_with_index.append((columns_to_retain.index(col.alias), col))
+
+        new_select_items = [y[1] for y in sorted(new_cols_with_index, key=lambda x: x[0])]
+
+        tds_row = LegendQLApiTdsRow.from_tds_frame("r", self.__base_frame)
+        for agg in self.__aggregates_list:
+            agg_sql_expr = agg[2].to_sql_expression({"r": new_query}, config)
+
+            new_select_items.append(
+                SingleColumn(alias=db_extension.quote_identifier(agg[0]), expression=agg_sql_expr)
+            )
+
+        new_query.select.selectItems = new_select_items
+        new_query.groupBy = [
+            (lambda x: x[c])(tds_row).to_sql_expression({"r": new_query}, config)
+            for c in self.__grouping_column_name_list
+        ]
+        return new_query
+
+    def to_pure(self, config: FrameToPureConfig) -> str:
+        group_strings = []
+        for col_name in self.__grouping_column_name_list:
+            group_strings.append(escape_column_name(col_name))
+
+        agg_strings = []
+        for agg in self.__aggregates_list:
+            map_expr_string = (agg[1].to_pure_expression(config) if isinstance(agg[1], PyLegendPrimitive)
+                               else convert_literal_to_literal_expression(agg[1]).to_pure_expression(config))
+            agg_expr_string = agg[2].to_pure_expression(config).replace(map_expr_string, "$c")
+            agg_strings.append(f"{escape_column_name(agg[0])}:{generate_pure_lambda('r', map_expr_string)}:"
+                               f"{generate_pure_lambda('c', agg_expr_string)}")
+
+        return (f"{self.__base_frame.to_pure(config)}{config.separator(1)}" +
+                f"->groupBy({config.separator(2)}"
+                f"~[{', '.join(group_strings)}],{config.separator(2, True)}"
+                f"~[{', '.join(agg_strings)}]{config.separator(1)}"
+                f")")
+
+    def base_frame(self) -> LegendQLApiBaseTdsFrame:
+        return self.__base_frame
+
+    def tds_frame_parameters(self) -> PyLegendList["LegendQLApiBaseTdsFrame"]:
+        return []
+
+    def calculate_columns(self) -> PyLegendSequence["TdsColumn"]:
+        base_columns = self.__base_frame.columns()
+        new_columns = []
+        for c in self.__grouping_column_name_list:
+            for base_col in base_columns:
+                if base_col.get_name() == c:
+                    new_columns.append(base_col.copy())
+                    break
+        for agg in self.__aggregates_list:
+            new_columns.append(tds_column_for_primitive(agg[0], agg[2]))
+        return new_columns
+
+    def validate(self) -> bool:
+        base_columns = self.__base_frame.columns()
+        for c in self.__grouping_column_name_list:
+            found_col = False
+            for base_col in base_columns:
+                if base_col.get_name() == c:
+                    found_col = True
+                    break
+            if not found_col:
+                raise ValueError(
+                    f"Column - '{c}' in group_by columns list doesn't exist in the current frame. "
+                    f"Current frame columns: {[x.get_name() for x in base_columns]}"
+                )
+
+        agg_cols = [c[0] for c in self.__aggregates_list]
+        new_cols = self.__grouping_column_name_list + agg_cols
+        if len(new_cols) == 0:
+            raise ValueError("At-least one grouping column or aggregate specification must be provided "
+                             "when using group_by function")
+        if len(new_cols) != len(set(new_cols)):
+            raise ValueError("Found duplicate column names in grouping columns and aggregation columns. "
+                             f"Grouping columns - {self.__grouping_column_name_list}, Aggregation columns - {agg_cols}")
+        return True

pylegend/core/tds/legendql_api/frames/functions/legendql_api_head_function.py

@@ -0,0 +1,74 @@
+# Copyright 2025 Goldman Sachs
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from pylegend._typing import (
+    PyLegendList,
+    PyLegendSequence,
+)
+from pylegend.core.tds.legendql_api.frames.legendql_api_applied_function_tds_frame import LegendQLApiAppliedFunction
+from pylegend.core.tds.sql_query_helpers import copy_query, create_sub_query
+from pylegend.core.sql.metamodel import (
+    QuerySpecification,
+    LongLiteral,
+)
+from pylegend.core.tds.tds_column import TdsColumn
+from pylegend.core.tds.tds_frame import FrameToSqlConfig
+from pylegend.core.tds.tds_frame import FrameToPureConfig
+from pylegend.core.tds.legendql_api.frames.legendql_api_base_tds_frame import LegendQLApiBaseTdsFrame
+
+
+__all__: PyLegendSequence[str] = [
+    "LegendQLApiHeadFunction"
+]
+
+
+class LegendQLApiHeadFunction(LegendQLApiAppliedFunction):
+    __base_frame: LegendQLApiBaseTdsFrame
+    __row_count: int
+
+    @classmethod
+    def name(cls) -> str:
+        return "head"
+
+    def __init__(self, base_frame: LegendQLApiBaseTdsFrame, row_count: int) -> None:
+        self.__base_frame = base_frame
+        self.__row_count = row_count
+
+    def to_sql(self, config: FrameToSqlConfig) -> QuerySpecification:
+        base_query = self.__base_frame.to_sql_query_object(config)
+        should_create_sub_query = (base_query.limit is not None)
+        new_query = (
+            create_sub_query(base_query, config, "root") if should_create_sub_query else
+            copy_query(base_query)
+        )
+        new_query.limit = LongLiteral(value=self.__row_count)
+        return new_query
+
+    def to_pure(self, config: FrameToPureConfig) -> str:
+        return (f"{self.__base_frame.to_pure(config)}{config.separator(1)}"
+                f"->limit({self.__row_count})")
+
+    def base_frame(self) -> LegendQLApiBaseTdsFrame:
+        return self.__base_frame
+
+    def tds_frame_parameters(self) -> PyLegendList["LegendQLApiBaseTdsFrame"]:
+        return []
+
+    def calculate_columns(self) -> PyLegendSequence["TdsColumn"]:
+        return [c.copy() for c in self.__base_frame.columns()]
+
+    def validate(self) -> bool:
+        if self.__row_count < 0:
+            raise ValueError("Row count argument of head/limit function cannot be negative")
+        return True

pylegend/core/tds/legendql_api/frames/functions/legendql_api_join_function.py

@@ -0,0 +1,214 @@
+# Copyright 2025 Goldman Sachs
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from pylegend._typing import (
+    PyLegendList,
+    PyLegendSequence,
+    PyLegendUnion,
+    PyLegendCallable,
+)
+from pylegend.core.language.legendql_api.legendql_api_tds_row import LegendQLApiTdsRow
+from pylegend.core.tds.legendql_api.frames.legendql_api_applied_function_tds_frame import LegendQLApiAppliedFunction
+from pylegend.core.tds.sql_query_helpers import copy_query, create_sub_query, extract_columns_for_subquery
+from pylegend.core.sql.metamodel import (
+    QuerySpecification,
+    Select,
+    SelectItem,
+    SingleColumn,
+    AliasedRelation,
+    TableSubquery,
+    Query,
+    Join,
+    JoinType,
+    JoinOn,
+    QualifiedNameReference,
+    QualifiedName,
+)
+from pylegend.core.tds.tds_column import TdsColumn
+from pylegend.core.tds.tds_frame import FrameToSqlConfig
+from pylegend.core.tds.tds_frame import FrameToPureConfig
+from pylegend.core.tds.legendql_api.frames.legendql_api_base_tds_frame import LegendQLApiBaseTdsFrame
+from pylegend.core.tds.legendql_api.frames.legendql_api_tds_frame import LegendQLApiTdsFrame
+from pylegend.core.language import (
+    PyLegendBoolean,
+    PyLegendBooleanLiteralExpression,
+    PyLegendPrimitive,
+    convert_literal_to_literal_expression,
+)
+from pylegend.core.language.shared.helpers import generate_pure_lambda
+
+
+__all__: PyLegendSequence[str] = [
+    "LegendQLApiJoinFunction"
+]
+
+
+class LegendQLApiJoinFunction(LegendQLApiAppliedFunction):
+    __base_frame: LegendQLApiBaseTdsFrame
+    __other_frame: LegendQLApiBaseTdsFrame
+    __join_condition: PyLegendCallable[[LegendQLApiTdsRow, LegendQLApiTdsRow], PyLegendUnion[bool, PyLegendBoolean]]
+    __join_type: str
+
+    @classmethod
+    def name(cls) -> str:
+        return "join"
+
+    def __init__(
+            self,
+            base_frame: LegendQLApiBaseTdsFrame,
+            other_frame: LegendQLApiTdsFrame,
+            join_condition: PyLegendCallable[[LegendQLApiTdsRow, LegendQLApiTdsRow], PyLegendUnion[bool, PyLegendBoolean]],
+            join_type: str
+    ) -> None:
+        self.__base_frame = base_frame
+        if not isinstance(other_frame, LegendQLApiBaseTdsFrame):
+            raise ValueError("Expected LegendQLApiBaseTdsFrame")  # pragma: no cover
+        self.__other_frame = other_frame
+        self.__join_condition = join_condition
+        self.__join_type = join_type
+
+    def to_sql(self, config: FrameToSqlConfig) -> QuerySpecification:
+        db_extension = config.sql_to_string_generator().get_db_extension()
+        base_query = copy_query(self.__base_frame.to_sql_query_object(config))
+        other_query = copy_query(self.__other_frame.to_sql_query_object(config))
+
+        join_type = (
+            JoinType.INNER if self.__join_type.lower() == 'inner' else (
+                JoinType.LEFT if self.__join_type.lower() in ('left_outer', 'leftouter') else
+                JoinType.RIGHT
+            )
+        )
+
+        left_row = LegendQLApiTdsRow.from_tds_frame('left', self.__base_frame)
+        right_row = LegendQLApiTdsRow.from_tds_frame('right', self.__other_frame)
+
+        join_expr = self.__join_condition(left_row, right_row)
+        if isinstance(join_expr, bool):
+            join_expr = PyLegendBoolean(PyLegendBooleanLiteralExpression(join_expr))
+        join_sql_expr = join_expr.to_sql_expression(
+            {
+                'left': create_sub_query(base_query, config, 'left'),
+                'right': create_sub_query(other_query, config, 'right'),
+            },
+            config
+        )
+
+        left_alias = db_extension.quote_identifier('left')
+        right_alias = db_extension.quote_identifier('right')
+        new_select_items: PyLegendList[SelectItem] = []
+        for c in self.__base_frame.columns():
+            q = db_extension.quote_identifier(c.get_name())
+            new_select_items.append(SingleColumn(q, QualifiedNameReference(name=QualifiedName(parts=[left_alias, q]))))
+        for c in self.__other_frame.columns():
+            q = db_extension.quote_identifier(c.get_name())
+            new_select_items.append(SingleColumn(q, QualifiedNameReference(name=QualifiedName(parts=[right_alias, q]))))
+
+        join_query = QuerySpecification(
+            select=Select(
+                selectItems=new_select_items,
+                distinct=False
+            ),
+            from_=[
+                Join(
+                    type_=join_type,
+                    left=AliasedRelation(
+                        relation=TableSubquery(query=Query(queryBody=base_query, limit=None, offset=None, orderBy=[])),
+                        alias=left_alias,
+                        columnNames=extract_columns_for_subquery(base_query)
+                    ),
+                    right=AliasedRelation(
+                        relation=TableSubquery(query=Query(queryBody=other_query, limit=None, offset=None, orderBy=[])),
+                        alias=right_alias,
+                        columnNames=extract_columns_for_subquery(other_query)
+                    ),
+                    criteria=JoinOn(expression=join_sql_expr)
+                )
+            ],
+            where=None,
+            groupBy=[],
+            having=None,
+            orderBy=[],
+            limit=None,
+            offset=None
+        )
+
+        wrapped_join_query = create_sub_query(join_query, config, "root")
+        return wrapped_join_query
+
+    def to_pure(self, config: FrameToPureConfig) -> str:
+        left_row = LegendQLApiTdsRow.from_tds_frame("l", self.__base_frame)
+        right_row = LegendQLApiTdsRow.from_tds_frame("r", self.__other_frame)
+        join_expr = self.__join_condition(left_row, right_row)
+        join_expr_string = (join_expr.to_pure_expression(config.push_indent(2))
+                            if isinstance(join_expr, PyLegendPrimitive) else
+                            convert_literal_to_literal_expression(join_expr).to_pure_expression(config.push_indent(2)))
+        join_kind = (
+            "INNER" if self.__join_type.lower() == 'inner' else
+            "LEFT" if self.__join_type.lower() in ('left_outer', 'leftouter') else
+            "RIGHT" if self.__join_type.lower() in ('right_outer', 'rightouter') else
+            "FULL"
+        )
+        return (f"{self.__base_frame.to_pure(config)}{config.separator(1)}" +
+                f"->join({config.separator(2)}"
+                f"{self.__other_frame.to_pure(config.push_indent(2))},{config.separator(2, True)}"
+                f"JoinKind.{join_kind},{config.separator(2, True)}"
+                f"{generate_pure_lambda('l, r', join_expr_string)}{config.separator(1)}"
+                f")")
+
+    def base_frame(self) -> LegendQLApiBaseTdsFrame:
+        return self.__base_frame
+
+    def tds_frame_parameters(self) -> PyLegendList["LegendQLApiBaseTdsFrame"]:
+        return [self.__other_frame]
+
+    def calculate_columns(self) -> PyLegendSequence["TdsColumn"]:
+        return (
+            [c.copy() for c in self.__base_frame.columns()] +
+            [c.copy() for c in self.__other_frame.columns()]
+        )
+
+    def validate(self) -> bool:
+        copy = self.__join_condition  # For MyPy
+        if not isinstance(copy, type(lambda x: 0)) or (copy.__code__.co_argcount != 2):
+            raise TypeError("Join condition function should be a lambda which takes two arguments (TDSRow, TDSRow)")
+
+        left_row = LegendQLApiTdsRow.from_tds_frame("left", self.__base_frame)
+        right_row = LegendQLApiTdsRow.from_tds_frame("right", self.__other_frame)
+
+        try:
+            result = self.__join_condition(left_row, right_row)
+        except Exception as e:
+            raise RuntimeError(
+                "Join condition function incompatible. Error occurred while evaluating. Message: " + str(e)
+            ) from e
+
+        if not isinstance(result, (bool, PyLegendBoolean)):
+            raise RuntimeError("Join condition function incompatible. Returns non boolean - " + str(type(result)))
+
+        left_cols = [c.get_name() for c in self.__base_frame.columns()]
+        right_cols = [c.get_name() for c in self.__other_frame.columns()]
+
+        final_cols = left_cols + right_cols
+        if len(final_cols) != len(set(final_cols)):
+            raise ValueError(
+                "Found duplicate columns in joined frames. Use rename function to ensure there are no duplicate columns "
+                f"in joined frames. Columns - Left Frame: {left_cols}, Right Frame: {right_cols}"
+            )
+
+        if self.__join_type.lower() not in ('inner', 'left_outer', 'right_outer', 'leftouter', 'rightouter'):
+            raise ValueError(
+                f"Unknown join type - {self.__join_type}. Supported types are - INNER, LEFT_OUTER, RIGHT_OUTER"
+            )
+
+        return True