Merge branch 'main' into telemetry

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

Author: saishreeeee

CHANGELOG.md

@@ -1,5 +1,14 @@
 # Release History
 
+# 4.0.4 (2025-06-16)
+
+- Update thrift client library after cleaning up unused fields and structs (databricks/databricks-sql-python#553 by @vikrantpuppala)
+- Refactor decimal conversion in PyArrow tables to use direct casting (databricks/databricks-sql-python#544 by @jayantsing-db)
+- Fix: `fetchall_arrow` to always return results in `arrow` format (databricks/databricks-sql-python#551 by @shivam2680)
+- Enhance cursor close handling and context manager exception management to prevent server side resource leaks (databricks/databricks-sql-python#554 by @madhav-db)
+- Added additional logging to enhance debugging (databricks/databricks-sql-python#556 by @saishreeeee)
+- Feature: Added support for complex data types such as Arrays and Map [Private Preview] (databricks/databricks-sql-python#559 by @jprakash-db)
+
 # 4.0.3 (2025-04-22)
 
 - Fix: Removed `packaging` dependency in favour of default libraries, for `urllib3` version checks (databricks/databricks-sql-python#547 by @jprakash-db)

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "databricks-sql-connector"
-version = "4.0.3"
+version = "4.0.4"
 description = "Databricks SQL Connector for Python"
 authors = ["Databricks <databricks-sql-connector-maintainers@databricks.com>"]
 license = "Apache-2.0"

src/databricks/sql/__init__.py

@@ -68,7 +68,7 @@ def __repr__(self):
 DATE = DBAPITypeObject("date")
 ROWID = DBAPITypeObject()
 
-__version__ = "4.0.3"
+__version__ = "4.0.4"
 USER_AGENT_NAME = "PyDatabricksSqlConnector"
 
 # These two functions are pyhive legacy

src/databricks/sql/parameters/__init__.py

@@ -12,4 +12,6 @@
     TimestampNTZParameter,
     TinyIntParameter,
     DecimalParameter,
+    MapParameter,
+    ArrayParameter,
 )

src/databricks/sql/parameters/native.py

@@ -1,12 +1,13 @@
 import datetime
 import decimal
 from enum import Enum, auto
-from typing import Optional, Sequence
+from typing import Optional, Sequence, Any
 
 from databricks.sql.exc import NotSupportedError
 from databricks.sql.thrift_api.TCLIService.ttypes import (
     TSparkParameter,
     TSparkParameterValue,
+    TSparkParameterValueArg,
 )
 
 import datetime
@@ -54,7 +55,17 @@ class DatabricksSupportedType(Enum):
 
 
 TAllowedParameterValue = Union[
-    str, int, float, datetime.datetime, datetime.date, bool, decimal.Decimal, None
+    str,
+    int,
+    float,
+    datetime.datetime,
+    datetime.date,
+    bool,
+    decimal.Decimal,
+    None,
+    list,
+    dict,
+    tuple,
 ]
 
 
@@ -82,6 +93,7 @@ class DbsqlParameterBase:
 
     CAST_EXPR: str
     name: Optional[str]
+    value: Any
 
     def as_tspark_param(self, named: bool) -> TSparkParameter:
         """Returns a TSparkParameter object that can be passed to the DBR thrift server."""
@@ -98,6 +110,10 @@ def as_tspark_param(self, named: bool) -> TSparkParameter:
     def _tspark_param_value(self):
         return TSparkParameterValue(stringValue=str(self.value))
 
+    def _tspark_value_arg(self):
+        """Returns a TSparkParameterValueArg object that can be passed to the DBR thrift server."""
+        return TSparkParameterValueArg(value=str(self.value), type=self._cast_expr())
+
     def _cast_expr(self):
         return self.CAST_EXPR
 
@@ -428,6 +444,99 @@ def __init__(self, value: int, name: Optional[str] = None):
     CAST_EXPR = DatabricksSupportedType.TINYINT.name
 
 
+class ArrayParameter(DbsqlParameterBase):
+    """Wrap a Python `Sequence` that will be bound to a Databricks SQL ARRAY type."""
+
+    def __init__(self, value: Sequence[Any], name: Optional[str] = None):
+        """
+        :value:
+            The value to bind for this parameter. This will be casted to a ARRAY.
+        :name:
+            If None, your query must contain a `?` marker. Like:
+
+            ```sql
+               SELECT * FROM table WHERE field = ?
+            ```
+            If not None, your query should contain a named parameter marker. Like:
+            ```sql
+                SELECT * FROM table WHERE field = :my_param
+            ```
+
+            The `name` argument to this function would be `my_param`.
+        """
+        self.name = name
+        self.value = [dbsql_parameter_from_primitive(val) for val in value]
+
+    def as_tspark_param(self, named: bool = False) -> TSparkParameter:
+        """Returns a TSparkParameter object that can be passed to the DBR thrift server."""
+
+        tsp = TSparkParameter(type=self._cast_expr())
+        tsp.arguments = [val._tspark_value_arg() for val in self.value]
+
+        if named:
+            tsp.name = self.name
+            tsp.ordinal = False
+        elif not named:
+            tsp.ordinal = True
+        return tsp
+
+    def _tspark_value_arg(self):
+        """Returns a TSparkParameterValueArg object that can be passed to the DBR thrift server."""
+        tva = TSparkParameterValueArg(type=self._cast_expr())
+        tva.arguments = [val._tspark_value_arg() for val in self.value]
+        return tva
+
+    CAST_EXPR = DatabricksSupportedType.ARRAY.name
+
+
+class MapParameter(DbsqlParameterBase):
+    """Wrap a Python `dict` that will be bound to a Databricks SQL MAP type."""
+
+    def __init__(self, value: dict, name: Optional[str] = None):
+        """
+        :value:
+            The value to bind for this parameter. This will be casted to a MAP.
+        :name:
+            If None, your query must contain a `?` marker. Like:
+
+            ```sql
+               SELECT * FROM table WHERE field = ?
+            ```
+            If not None, your query should contain a named parameter marker. Like:
+            ```sql
+                SELECT * FROM table WHERE field = :my_param
+            ```
+
+            The `name` argument to this function would be `my_param`.
+        """
+        self.name = name
+        self.value = [
+            dbsql_parameter_from_primitive(item)
+            for key, val in value.items()
+            for item in (key, val)
+        ]
+
+    def as_tspark_param(self, named: bool = False) -> TSparkParameter:
+        """Returns a TSparkParameter object that can be passed to the DBR thrift server."""
+
+        tsp = TSparkParameter(type=self._cast_expr())
+        tsp.arguments = [val._tspark_value_arg() for val in self.value]
+        if named:
+            tsp.name = self.name
+            tsp.ordinal = False
+        elif not named:
+            tsp.ordinal = True
+        return tsp
+
+    def _tspark_value_arg(self):
+        """Returns a TSparkParameterValueArg object that can be passed to the DBR thrift server."""
+        tva = TSparkParameterValueArg(type=self._cast_expr())
+        tva.arguments = [val._tspark_value_arg() for val in self.value]
+        return tva
+
+    CAST_EXPR = DatabricksSupportedType.MAP.name
+
+
 class DecimalParameter(DbsqlParameterBase):
     """Wrap a Python `Decimal` that will be bound to a Databricks SQL DECIMAL type."""
 
@@ -543,23 +652,26 @@ def dbsql_parameter_from_primitive(
     # havoc. We can't use TYPE_INFERRENCE_MAP because mypy doesn't trust
     # its logic
 
-    if type(value) is int:
+    if isinstance(value, bool):
+        return BooleanParameter(value=value, name=name)
+    elif isinstance(value, int):
         return dbsql_parameter_from_int(value, name=name)
-    elif type(value) is str:
+    elif isinstance(value, str):
         return StringParameter(value=value, name=name)
-    elif type(value) is float:
+    elif isinstance(value, float):
         return FloatParameter(value=value, name=name)
-    elif type(value) is datetime.datetime:
+    elif isinstance(value, datetime.datetime):
         return TimestampParameter(value=value, name=name)
-    elif type(value) is datetime.date:
+    elif isinstance(value, datetime.date):
         return DateParameter(value=value, name=name)
-    elif type(value) is bool:
-        return BooleanParameter(value=value, name=name)
-    elif type(value) is decimal.Decimal:
+    elif isinstance(value, decimal.Decimal):
         return DecimalParameter(value=value, name=name)
+    elif isinstance(value, dict):
+        return MapParameter(value=value, name=name)
+    elif isinstance(value, Sequence) and not isinstance(value, str):
+        return ArrayParameter(value=value, name=name)
     elif value is None:
         return VoidParameter(value=value, name=name)
-
     else:
         raise NotSupportedError(
             f"Could not infer parameter type from value: {value} - {type(value)} \n"
@@ -581,6 +693,8 @@ def dbsql_parameter_from_primitive(
     TimestampNTZParameter,
     TinyIntParameter,
     DecimalParameter,
+    ArrayParameter,
+    MapParameter,
 ]
 
 

src/databricks/sql/utils.py

@@ -5,10 +5,10 @@
 import decimal
 from abc import ABC, abstractmethod
 from collections import OrderedDict, namedtuple
-from collections.abc import Iterable
+from collections.abc import Mapping
 from decimal import Decimal
 from enum import Enum
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional, Union, Sequence
 import re
 
 import lz4.frame
@@ -429,7 +429,7 @@ def user_friendly_error_message(self, no_retry_reason, attempt, elapsed):
 # Taken from PyHive
 class ParamEscaper:
     _DATE_FORMAT = "%Y-%m-%d"
-    _TIME_FORMAT = "%H:%M:%S.%f"
+    _TIME_FORMAT = "%H:%M:%S.%f %z"
     _DATETIME_FORMAT = "{} {}".format(_DATE_FORMAT, _TIME_FORMAT)
 
     def escape_args(self, parameters):
@@ -458,13 +458,22 @@ def escape_string(self, item):
         return "'{}'".format(item.replace("\\", "\\\\").replace("'", "\\'"))
 
     def escape_sequence(self, item):
-        l = map(str, map(self.escape_item, item))
-        return "(" + ",".join(l) + ")"
+        l = map(self.escape_item, item)
+        l = list(map(str, l))
+        return "ARRAY(" + ",".join(l) + ")"
+
+    def escape_mapping(self, item):
+        l = map(
+            self.escape_item,
+            (element for key, value in item.items() for element in (key, value)),
+        )
+        l = list(map(str, l))
+        return "MAP(" + ",".join(l) + ")"
 
     def escape_datetime(self, item, format, cutoff=0):
         dt_str = item.strftime(format)
         formatted = dt_str[:-cutoff] if cutoff and format.endswith(".%f") else dt_str
-        return "'{}'".format(formatted)
+        return "'{}'".format(formatted.strip())
 
     def escape_decimal(self, item):
         return str(item)
@@ -476,14 +485,16 @@ def escape_item(self, item):
             return self.escape_number(item)
         elif isinstance(item, str):
             return self.escape_string(item)
-        elif isinstance(item, Iterable):
-            return self.escape_sequence(item)
         elif isinstance(item, datetime.datetime):
             return self.escape_datetime(item, self._DATETIME_FORMAT)
         elif isinstance(item, datetime.date):
             return self.escape_datetime(item, self._DATE_FORMAT)
         elif isinstance(item, decimal.Decimal):
             return self.escape_decimal(item)
+        elif isinstance(item, Sequence):
+            return self.escape_sequence(item)
+        elif isinstance(item, Mapping):
+            return self.escape_mapping(item)
         else:
             raise exc.ProgrammingError("Unsupported object {}".format(item))
 

tests/e2e/test_complex_types.py

@@ -1,5 +1,6 @@
 import pytest
 from numpy import ndarray
+from typing import Sequence
 
 from tests.e2e.test_driver import PySQLPytestTestCase
 
@@ -17,8 +18,11 @@ def table_fixture(self, connection_details):
                 CREATE TABLE IF NOT EXISTS pysql_test_complex_types_table (
                     array_col ARRAY<STRING>,
                     map_col MAP<STRING, INTEGER>,
-                    struct_col STRUCT<field1: STRING, field2: INTEGER>
-                )
+                    struct_col STRUCT<field1: STRING, field2: INTEGER>,
+                    array_array_col ARRAY<ARRAY<STRING>>,
+                    array_map_col ARRAY<MAP<STRING, INTEGER>>,
+                    map_array_col MAP<STRING, ARRAY<STRING>>
+                ) USING DELTA
                 """
             )
             # Insert a record
@@ -28,17 +32,27 @@ def table_fixture(self, connection_details):
                 VALUES (
                     ARRAY('a', 'b', 'c'),
                     MAP('a', 1, 'b', 2, 'c', 3),
-                    NAMED_STRUCT('field1', 'a', 'field2', 1)
+                    NAMED_STRUCT('field1', 'a', 'field2', 1),
+                    ARRAY(ARRAY('a','b','c')),
+                    ARRAY(MAP('a', 1, 'b', 2, 'c', 3)),
+                    MAP('a', ARRAY('a', 'b', 'c'), 'b', ARRAY('d', 'e'))
                 )
                 """
             )
             yield
             # Clean up the table after the test
-            cursor.execute("DROP TABLE IF EXISTS pysql_test_complex_types_table")
+            cursor.execute("DELETE FROM pysql_test_complex_types_table")
 
     @pytest.mark.parametrize(
         "field,expected_type",
-        [("array_col", ndarray), ("map_col", list), ("struct_col", dict)],
+        [
+            ("array_col", ndarray),
+            ("map_col", list),
+            ("struct_col", dict),
+            ("array_array_col", ndarray),
+            ("array_map_col", ndarray),
+            ("map_array_col", list),
+        ],
     )
     def test_read_complex_types_as_arrow(self, field, expected_type, table_fixture):
         """Confirms the return types of a complex type field when reading as arrow"""
@@ -50,7 +64,17 @@ def test_read_complex_types_as_arrow(self, field, expected_type, table_fixture):
 
         assert isinstance(result[field], expected_type)
 
-    @pytest.mark.parametrize("field", [("array_col"), ("map_col"), ("struct_col")])
+    @pytest.mark.parametrize(
+        "field",
+        [
+            ("array_col"),
+            ("map_col"),
+            ("struct_col"),
+            ("array_array_col"),
+            ("array_map_col"),
+            ("map_array_col"),
+        ],
+    )
     def test_read_complex_types_as_string(self, field, table_fixture):
         """Confirms the return type of a complex type that is returned as a string"""
         with self.cursor(

tests/e2e/test_driver.py

@@ -856,7 +856,9 @@ def test_closing_a_closed_connection_doesnt_fail(self, caplog):
                     raise KeyboardInterrupt("Simulated interrupt")
         finally:
             if conn is not None:
-                assert not conn.open, "Connection should be closed after KeyboardInterrupt"
+                assert (
+                    not conn.open
+                ), "Connection should be closed after KeyboardInterrupt"
 
     def test_cursor_close_properly_closes_operation(self):
         """Test that Cursor.close() properly closes the active operation handle on the server."""
@@ -883,7 +885,9 @@ def test_cursor_close_properly_closes_operation(self):
                         raise KeyboardInterrupt("Simulated interrupt")
         finally:
             if cursor is not None:
-                assert not cursor.open, "Cursor should be closed after KeyboardInterrupt"
+                assert (
+                    not cursor.open
+                ), "Cursor should be closed after KeyboardInterrupt"
 
     def test_nested_cursor_context_managers(self):
         """Test that nested cursor context managers properly close operations on the server."""