Merge branch 'expr-typing' of https://github.com/OutSquareCapital/duckdb-python into expr-typing

Author: OutSquareCapital

duckdb_packaging/setuptools_scm_version.py

@@ -12,7 +12,7 @@
 from ._versioning import format_version, parse_version
 
 # MAIN_BRANCH_VERSIONING should be 'True' on main branch only
-MAIN_BRANCH_VERSIONING = True
+MAIN_BRANCH_VERSIONING = False
 
 SCM_PRETEND_ENV_VAR = "SETUPTOOLS_SCM_PRETEND_VERSION_FOR_DUCKDB"
 SCM_GLOBAL_PRETEND_ENV_VAR = "SETUPTOOLS_SCM_PRETEND_VERSION"

external/duckdb

@@ -256,7 +256,9 @@ test = [ # dependencies used for running tests
     "pandas<3.0.0; python_version < '3.11'",
     "pyarrow>=23.0.0; python_version >= '3.10' and (sys_platform != 'win32' or platform_machine != 'ARM64')",
     "pyarrow>=18.0.0; python_version < '3.10' and (sys_platform != 'win32' or platform_machine != 'ARM64')",
-    "torch>=2.2.2; python_version < '3.14' and ( sys_platform != 'darwin' or platform_machine != 'x86_64' or python_version < '3.13' ) and ( sys_platform != 'win32' or platform_machine != 'ARM64' or python_version > '3.11' )",
+    "torch>=2.2.2; python_version < '3.13' and ( sys_platform != 'win32' or platform_machine != 'ARM64' or python_version > '3.11' )",
+    "torch>=2.2.2,<2.11.0; python_version >= '3.13' and python_version < '3.14' and ( sys_platform != 'darwin' or platform_machine != 'x86_64' ) and ( sys_platform != 'win32' or platform_machine != 'ARM64' )",
+    "torch>=2.10.0; python_version >= '3.14' and ( sys_platform != 'darwin' or platform_machine != 'x86_64' ) and ( sys_platform != 'win32' or platform_machine != 'ARM64' )",
     "tensorflow==2.14.0; sys_platform == 'darwin' and python_version < '3.12'",
     "tensorflow-cpu>=2.14.0; sys_platform == 'linux' and platform_machine != 'aarch64' and python_version < '3.12'",
     "tensorflow-cpu>=2.14.0; sys_platform == 'win32' and platform_machine != 'ARM64' and python_version < '3.12'",

pyproject.toml

@@ -444,6 +444,7 @@ static bool KeyIsHashable(const LogicalType &type) {
 	case LogicalTypeId::DATE:
 	case LogicalTypeId::UUID:
 	case LogicalTypeId::INTERVAL:
+	case LogicalTypeId::GEOMETRY:
 		return true;
 	case LogicalTypeId::LIST:
 	case LogicalTypeId::ARRAY:
@@ -511,6 +512,7 @@ py::object PythonObject::FromValue(const Value &val, const LogicalType &type,
 	case LogicalTypeId::VARCHAR:
 		return py::cast(StringValue::Get(val));
 	case LogicalTypeId::BLOB:
+	case LogicalTypeId::GEOMETRY:
 		return py::bytes(StringValue::Get(val));
 	case LogicalTypeId::BIT:
 		return py::cast(Bit::ToString(StringValue::Get(val)));

src/duckdb_py/native/python_objects.cpp

@@ -682,6 +682,7 @@ void ArrayWrapper::Append(idx_t current_offset, Vector &input, idx_t source_size
 		may_have_null = ConvertColumn<string_t, PyObject *, duckdb_py_convert::StringConvert>(append_data);
 		break;
 	case LogicalTypeId::BLOB:
+	case LogicalTypeId::GEOMETRY:
 		may_have_null = ConvertColumn<string_t, PyObject *, duckdb_py_convert::BlobConvert>(append_data);
 		break;
 	case LogicalTypeId::BIT:

src/duckdb_py/numpy/array_wrapper.cpp

@@ -61,6 +61,7 @@ static idx_t GetNumpyTypeWidth(const LogicalType &type) {
 	case LogicalTypeId::UUID:
 	case LogicalTypeId::ARRAY:
 	case LogicalTypeId::VARIANT:
+	case LogicalTypeId::GEOMETRY:
 		return sizeof(PyObject *);
 	default:
 		throw NotImplementedException("Unsupported type \"%s\" for DuckDB -> NumPy conversion", type.ToString());
@@ -124,6 +125,7 @@ string RawArrayWrapper::DuckDBToNumpyDtype(const LogicalType &type) {
 	case LogicalTypeId::UUID:
 	case LogicalTypeId::ARRAY:
 	case LogicalTypeId::VARIANT:
+	case LogicalTypeId::GEOMETRY:
 		return "object";
 	case LogicalTypeId::ENUM: {
 		auto size = EnumType::GetSize(type);

src/duckdb_py/numpy/raw_array_wrapper.cpp

@@ -969,8 +969,7 @@ duckdb::pyarrow::Table DuckDBPyRelation::ToArrowTableInternal(idx_t batch_size,
 		ScopedConfigSetting scoped_setting(
 		    config,
 		    [&batch_size](ClientConfig &config) {
-			    config.get_result_collector = [&batch_size](ClientContext &context,
-			                                                PreparedStatementData &data) -> PhysicalOperator & {
+			    config.get_result_collector = [&batch_size](ClientContext &context, PreparedStatementData &data) {
 				    return PhysicalArrowCollector::Create(context, data, batch_size);
 			    };
 		    },
@@ -999,8 +998,7 @@ py::object DuckDBPyRelation::ToArrowCapsule(const py::object &requested_schema)
 		ScopedConfigSetting scoped_setting(
 		    config,
 		    [&batch_size](ClientConfig &config) {
-			    config.get_result_collector = [&batch_size](ClientContext &context,
-			                                                PreparedStatementData &data) -> PhysicalOperator & {
+			    config.get_result_collector = [&batch_size](ClientContext &context, PreparedStatementData &data) {
 				    return PhysicalArrowCollector::Create(context, data, batch_size);
 			    };
 		    },

src/duckdb_py/pyrelation.cpp

@@ -0,0 +1,30 @@
+import pandas as pd
+
+
+class TestGeometry:
+    def test_fetchall(self, duckdb_cursor):
+        duckdb_cursor.execute("SELECT 'POINT(1 2)'::GEOMETRY AS geom")
+        results = duckdb_cursor.fetchall()
+        assert isinstance(results[0][0], bytes)
+
+    def test_fetchnumpy(self, duckdb_cursor):
+        duckdb_cursor.execute("SELECT 'POINT(1 2)'::GEOMETRY AS geom")
+        results = duckdb_cursor.fetchnumpy()
+        assert isinstance(results["geom"][0], (bytes, bytearray))
+
+    def test_df(self, duckdb_cursor):
+        duckdb_cursor.execute("SELECT 'POINT(1 2)'::GEOMETRY AS geom")
+        df = duckdb_cursor.df()
+        assert isinstance(df["geom"].iloc[0], (bytes, bytearray))
+
+    def test_null(self, duckdb_cursor):
+        duckdb_cursor.execute("SELECT NULL::GEOMETRY AS geom")
+        results = duckdb_cursor.fetchall()
+        assert results[0][0] is None
+
+    def test_multiple_rows(self, duckdb_cursor):
+        duckdb_cursor.execute("SELECT 'POINT(1 2)'::GEOMETRY AS geom UNION ALL SELECT NULL::GEOMETRY")
+        df = duckdb_cursor.df()
+        assert df.shape[0] == 2
+        assert isinstance(df["geom"].iloc[0], (bytes, bytearray))
+        assert pd.isna(df["geom"].iloc[1])