OIiveirra commented on code in PR #69032:
URL: https://github.com/apache/doris/pull/69032#discussion_r4376007507
##########
be/src/exprs/function/geo/functions_geo.cpp:
##########
@@ -18,32 +18,869 @@
#include "exprs/function/geo/functions_geo.h"
#include <glog/logging.h>
+#include <s2/mutable_s2shape_index.h>
+#include <s2/s2boolean_operation.h>
+#include <s2/s2builderutil_s2polyline_vector_layer.h>
+#include <s2/s2closest_edge_query.h>
+#include <s2/s2closest_point_query.h>
+#include <s2/s2earth.h>
+#include <s2/s2lax_polygon_shape.h>
+#include <s2/s2loop.h>
#include <s2/s2point.h>
+#include <s2/s2point_index.h>
+#include <s2/s2point_vector_shape.h>
+#include <s2/s2polygon.h>
+#include <s2/s2polyline.h>
+#include <s2/s2predicates.h>
#include <algorithm>
+#include <array>
+#include <bit>
#include <boost/iterator/iterator_facade.hpp>
+#include <boost/multiprecision/cpp_int.hpp>
+#include <cmath>
+#include <cstdint>
+#include <limits>
+#include <map>
+#include <optional>
+#include <set>
#include <utility>
+#include "common/cast_set.h"
#include "common/compiler_util.h"
#include "core/assert_cast.h"
#include "core/block/block.h"
#include "core/block/column_with_type_and_name.h"
#include "core/column/column.h"
#include "core/column/column_array.h"
#include "core/column/column_execute_util.h"
+#include "core/column/column_map.h"
#include "core/column/column_nullable.h"
+#include "core/column/column_struct.h"
+#include "core/column/column_varbinary.h"
+#include "core/custom_allocator.h"
#include "core/data_type/data_type_array.h"
+#include "core/data_type/data_type_map.h"
#include "core/data_type/data_type_nullable.h"
#include "core/data_type/data_type_number.h"
+#include "core/data_type/data_type_spatial.h"
#include "core/data_type/data_type_string.h"
+#include "core/data_type/data_type_struct.h"
+#include "core/data_type/data_type_varbinary.h"
#include "core/data_type/define_primitive_type.h"
#include "core/string_ref.h"
#include "exprs/function/geo/geo_common.h"
#include "exprs/function/geo/geo_types.h"
+#include "exprs/function/geo/wkb_parse.h"
#include "exprs/function/simple_function_factory.h"
+#include "exprs/function/string_hex_util.h"
namespace doris {
+static bool is_spatial_type(const DataTypePtr& type) {
+ const auto primitive_type = remove_nullable(type)->get_primitive_type();
+ return primitive_type == TYPE_GEOMETRY || primitive_type == TYPE_GEOGRAPHY;
+}
+
+static Status validate_geography_semantics(const DataTypePtr& type, const
char* function_name) {
+ const auto& nested_type = remove_nullable(type);
+ if (!is_spatial_type(nested_type)) {
+ return Status::OK();
+ }
+
+ const auto* spatial_type = dynamic_cast<const
DataTypeSpatial*>(nested_type.get());
+ DCHECK(spatial_type != nullptr);
+ if (spatial_type->get_primitive_type() == TYPE_GEOGRAPHY &&
+ spatial_type->crs() == "OGC:CRS84" && spatial_type->algorithm() ==
"spherical") {
+ return Status::OK();
+ }
+ return Status::NotSupported(
+ "Function {} requires GEOGRAPHY(OGC:CRS84, spherical) for spatial
inputs",
+ function_name);
+}
+
+static std::unique_ptr<GeoShape> decode_geo_shape(StringRef value, const
DataTypePtr& type) {
+ if (!is_spatial_type(type)) {
+ return GeoShape::from_encoded(value.data, value.size);
+ }
+
+ GeoParseStatus status;
+ auto shape = GeoShape::from_wkb_bytes(value.data, value.size, status);
+ return status == GeoParseStatus::GEO_PARSE_OK ? std::move(shape) : nullptr;
+}
+
+static bool any_geo_argument_is_null(const Block& block, const ColumnNumbers&
arguments,
+ size_t row) {
+ return std::ranges::any_of(arguments, [&](auto argument) {
+ return block.get_by_position(argument).column->is_null_at(row);
+ });
+}
+
+static Status decode_calculation_shape(StringRef value, const DataTypePtr&
type,
+ std::unique_ptr<GeoShape>* shape,
+ const char* function_name) {
+ *shape = decode_geo_shape(value, type);
+ // POLYGON EMPTY decodes successfully, but distance helpers retain
infinity/max sentinels.
+ // Reject it for typed calculations just like empty points/lines, leaving
raw WKB untouched.
+ if (is_spatial_type(type) &&
+ (*shape == nullptr ||
+ ((*shape)->type() == GeoShapeType::GEO_SHAPE_POLYGON &&
+ assert_cast<const GeoPolygon&>(**shape).polygon()->num_loops() ==
0))) {
+ return Status::NotSupported("Function {} does not support this
Geography WKB shape",
+ function_name);
+ }
+ return Status::OK();
+}
+
+// Calculations use canonical positions; raw spatial accessors still return
the original WKB.
+static S2Point canonical_geography_point(const S2Point& point) {
+ const auto coordinate = S2LatLng(point);
+ const double latitude = coordinate.lat().radians();
+ // All longitudes at a pole and +/-180 at the date line denote the same
position.
+ if (std::abs(latitude) == M_PI_2) {
+ return {0, 0, latitude > 0 ? 1.0 : -1.0};
+ }
+ if (coordinate.lng().radians() == M_PI) {
+ // Flip only sin(longitude): rebuilding latitude can change its last
bit.
+ return {point.x(), -point.y(), point.z()};
+ }
+ return point;
+}
+
+// Error-free TwoSum preserves tiny differences from a half-turn (e.g. -1e-15
+ 180).
+static std::pair<double, double> geography_degree_sum(double a, double b) {
+ const double high = a + b;
+ const double offset = high - a;
+ return {high, (a - (high - offset)) + (b - offset)};
+}
+
+// Test coplanarity of the original degree ordinates, not independently
rounded S2 vectors.
+// For example (45,-45), (90,0), (135,45) share an arc although ToPoint's
determinant is nonzero.
+static bool geography_common_great_circle(const GeoCoordinate& a, const
GeoCoordinate& b,
+ const GeoCoordinate& c) {
+ using boost::multiprecision::cpp_int;
+ const std::array<double, 6> ordinates {a.y, a.x, b.y, b.x, c.y, c.x};
+ std::array<uint64_t, 6> significands;
+ std::array<int, 6> exponents;
+ int denominator_bits = 0;
+ for (size_t i = 0; i < ordinates.size(); ++i) {
+ const auto bits = std::bit_cast<uint64_t>(ordinates[i]);
+ const int biased = (bits >> 52) & 0x7ff;
+ uint64_t significand = bits & ((uint64_t {1} << 52) - 1);
+ int exponent = biased ? biased - 1075 : -1074;
+ if (biased) {
+ significand |= uint64_t {1} << 52;
+ }
+ if (significand != 0) {
+ const int zeros = std::countr_zero(significand);
+ significand >>= zeros;
+ exponent += zeros;
+ denominator_bits = std::max(denominator_bits, -exponent);
+ } else {
+ exponent = 0;
+ }
+ significands[i] = significand;
+ exponents[i] = exponent;
+ }
+ std::array<cpp_int, 6> angles;
+ for (size_t i = 0; i < angles.size(); ++i) {
+ angles[i] = cpp_int(significands[i]) << (denominator_bits +
exponents[i]);
+ if (std::signbit(ordinates[i])) {
+ angles[i] = -angles[i];
+ }
+ }
+ // Binary-rational degrees have denominator 2^q. Exp(i*angle) is a power
of a
+ // root of unity of order 360*2^q. Its minimal polynomial is Phi_90(t),
with
+ // t = z^(2^(q+2)), and Phi_90(t) = t^24+t^21-t^15-t^12-t^9+t^3+1.
+ // Expand cos(lat_i)cos(lat_j)sin(lat_k)sin(lon_j-lon_i), cyclically, into
+ // 48 integer-coefficient powers. Equal residues modulo 2^(q+2) share one
+ // polynomial; reduction by Phi_90 proves zero exactly, including tiny
nonzero controls.
+ const int shift = denominator_bits + 2;
+ const cpp_int basis = cpp_int(1) << shift;
+ const cpp_int period = 90 * basis;
+ std::map<cpp_int, std::array<int, 90>> polynomials;
+ for (int i = 0; i < 3; ++i) {
+ const int j = (i + 1) % 3;
+ const int k = (i + 2) % 3;
+ for (int mask = 0; mask < 16; ++mask) {
+ std::array<int, 4> signs;
+ for (int bit = 0; bit < 4; ++bit) {
+ signs[bit] = (mask & (1 << bit)) ? 1 : -1;
+ }
+ cpp_int power = signs[0] * angles[2 * i] + signs[1] * angles[2 *
j] +
+ signs[2] * angles[2 * k] +
+ signs[3] * (angles[2 * j + 1] - angles[2 * i + 1]);
+ power %= period;
+ if (power < 0) {
+ power += period;
+ }
+ const int quotient = (power >> shift).convert_to<int>();
+ const cpp_int residue = power - (cpp_int(quotient) << shift);
+ polynomials[residue][quotient] += signs[2] * signs[3];
+ }
+ }
+ for (auto& [residue, coefficients] : polynomials) {
+ for (int i = 89; i >= 24; --i) {
+ const int coefficient = coefficients[i];
+ coefficients[i] = 0;
+ coefficients[i - 3] -= coefficient;
+ coefficients[i - 9] += coefficient;
+ coefficients[i - 12] += coefficient;
+ coefficients[i - 15] += coefficient;
+ coefficients[i - 21] -= coefficient;
+ coefficients[i - 24] -= coefficient;
+ }
+ if (std::ranges::any_of(coefficients, [](int value) { return value !=
0; })) {
+ return false;
+ }
+ }
+ return true;
+}
+
+// Keep each atomic segment in its own shape. Pinned S2 ProcessEdge1 uses a
chain's
+// global start as a local offset in one endpoint path, so multiple chains per
line
+// shape are unsafe. Closed AA shapes restore internal vertices in the OPEN
model.
+class GeographyLineShape final : public S2Shape {
+public:
+ GeographyLineShape(const S2Point& a, const S2Point& b) {
_edge.emplace_back(a, b); }
+ int num_edges() const override { return 1; }
+ Edge edge(int i) const override {
+ DCHECK_EQ(i, 0);
+ return _edge.front();
+ }
+ int dimension() const override { return 1; }
+ ReferencePoint GetReferencePoint() const override { return
ReferencePoint::Contained(false); }
+ int num_chains() const override { return 1; }
+ Chain chain(int i) const override {
+ DCHECK_EQ(i, 0);
+ return {0, 1};
+ }
+ Edge chain_edge(int i, int j) const override {
+ DCHECK_EQ(i, 0);
+ return edge(j);
+ }
+ ChainPosition chain_position(int i) const override {
+ DCHECK_EQ(i, 0);
+ return {0, 0};
+ }
+
+private:
+ // Owning storage is checked/accounted by the current query's allocator.
+ DorisVector<Edge> _edge;
+};
+
+class GeographyEdgeNodes {
+public:
+ GeographyEdgeNodes(StringRef left, StringRef right) {
+ GeoCoordinateList coordinates;
+ const auto left_status = WkbParse::xy_coordinates(left.data,
left.size, &coordinates);
+ const auto right_status = WkbParse::xy_coordinates(right.data,
right.size, &coordinates);
+ DCHECK_EQ(left_status, GeoParseStatus::GEO_PARSE_OK);
+ DCHECK_EQ(right_status, GeoParseStatus::GEO_PARSE_OK);
+ for (auto coordinate : coordinates.list) {
+ if (coordinate.x == 180) {
+ coordinate.x = -180;
+ }
+ const auto point = canonical_geography_point(
+ S2LatLng::FromDegrees(coordinate.y,
coordinate.x).ToPoint());
+ if (!_coordinates.emplace(point, coordinate).second) {
+ continue;
+ }
+ _vertex_index.Add(point, 0);
+ if (std::abs(coordinate.y) != 90) {
+ _meridians[meridian(coordinate.x)][coordinate.x < 0].push_back(
+ {coordinate.y, point});
+ }
+ }
+ for (auto& entry : _meridians) {
+ for (auto& side : entry.second) {
+ side.push_back({-90, {0, 0, -1}});
+ side.push_back({90, {0, 0, 1}});
+ std::ranges::sort(side, {}, &Vertex::latitude);
+ }
+ }
+ }
+
+ void append_edge(const S2Point& raw_start, const S2Point& raw_end,
+ DorisVector<S2Point>* output) const {
+ const auto start = canonical_geography_point(raw_start);
+ const auto end = canonical_geography_point(raw_end);
+ output->push_back(start);
+ if (start == end) {
+ return;
+ }
+ if (meridian_edge(start, end, [&](const auto& vertices, double from,
double to) {
+ append_range(vertices, from, to, output);
+ if (std::abs(to) == 90) {
+ const S2Point pole(0, 0, to > 0 ? 1 : -1);
+ if (end != pole) {
+ output->push_back(pole);
+ }
+ }
+ })) {
+ return;
+ }
+ const auto& a = _coordinates.at(start);
+ const auto& b = _coordinates.at(end);
+ const bool wraps = geography_degree_sum(std::max(a.x, b.x),
-std::min(a.x, b.x)) >
+ std::pair {180.0, 0.0};
+ const bool east = (a.x < b.x) != wraps;
+ S2ClosestPointQuery<int> query(&_vertex_index);
+ const double conversion_error = 8 *
std::numeric_limits<double>::epsilon();
+ // Near antipodes, endpoint rounding rotates the edge plane by
O(epsilon/sin(length)).
+ // Widen candidate search only; the exact raw-angle proof still
decides membership.
+ // Every point on a minor arc is at most pi/2 from one endpoint,
including when
+ // rounded antipodes make the computed normal vanish.
+ const double search_radius =
+ start.DotProd(end) >= 0
+ ? conversion_error
+ : std::min(M_PI_2 + conversion_error,
+ 4 * conversion_error /
start.CrossProd(end).Norm());
+
query.mutable_options()->set_conservative_max_distance(S1Angle::Radians(search_radius));
+ S2ClosestPointQuery<int>::EdgeTarget target(start, end);
+ DorisVector<S2Point> on_edge;
+ for (const auto& candidate : query.FindClosestPoints(&target)) {
+ const auto& point = candidate.point();
+ const auto& c = _coordinates.at(point);
+ // Longitude is strictly monotone along every non-meridian minor
arc.
+ // Check the original interval (including date-line wrap), not a
distance
+ // tolerance that could append a point just beyond an endpoint.
+ const bool within = wraps ? (c.x < std::min(a.x, b.x) || c.x >
std::max(a.x, b.x))
+ : (c.x > std::min(a.x, b.x) && c.x <
std::max(a.x, b.x));
+ if (within && geography_common_great_circle(a, b, c)) {
Review Comment:
Fixed in c7d31ed3989, retained in af66f59ab50. Same-circle coverage is
merged before noding, with error-bucketed candidate lookup. Added proof-count,
mixed-circle and boundary regressions; 115 geo/WKB UTs and both targeted suites
passed.
##########
be/src/exprs/function/geo/functions_geo.cpp:
##########
@@ -32,18 +44,302 @@
#include "core/column/column_array.h"
#include "core/column/column_execute_util.h"
#include "core/column/column_nullable.h"
+#include "core/column/column_varbinary.h"
#include "core/data_type/data_type_array.h"
#include "core/data_type/data_type_nullable.h"
#include "core/data_type/data_type_number.h"
+#include "core/data_type/data_type_spatial.h"
#include "core/data_type/data_type_string.h"
+#include "core/data_type/data_type_varbinary.h"
#include "core/data_type/define_primitive_type.h"
#include "core/string_ref.h"
#include "exprs/function/geo/geo_common.h"
#include "exprs/function/geo/geo_types.h"
+#include "exprs/function/geo/wkb_parse.h"
#include "exprs/function/simple_function_factory.h"
+#include "exprs/function/string_hex_util.h"
namespace doris {
+static bool is_spatial_type(const DataTypePtr& type) {
+ const auto primitive_type = remove_nullable(type)->get_primitive_type();
+ return primitive_type == TYPE_GEOMETRY || primitive_type == TYPE_GEOGRAPHY;
+}
+
+static Status validate_geography_semantics(const DataTypePtr& type, const
char* function_name) {
+ const auto& nested_type = remove_nullable(type);
+ if (!is_spatial_type(nested_type)) {
+ return Status::OK();
+ }
+
+ const auto* spatial_type = dynamic_cast<const
DataTypeSpatial*>(nested_type.get());
+ DCHECK(spatial_type != nullptr);
+ if (spatial_type->get_primitive_type() == TYPE_GEOGRAPHY &&
+ spatial_type->crs() == "OGC:CRS84" && spatial_type->algorithm() ==
"spherical") {
+ return Status::OK();
+ }
+ return Status::NotSupported(
+ "Function {} requires GEOGRAPHY(OGC:CRS84, spherical) for spatial
inputs",
+ function_name);
+}
+
+static std::unique_ptr<GeoShape> decode_geo_shape(StringRef value, const
DataTypePtr& type) {
+ if (!is_spatial_type(type)) {
+ return GeoShape::from_encoded(value.data, value.size);
+ }
+
+ GeoParseStatus status;
+ auto shape = GeoShape::from_wkb_bytes(value.data, value.size, status);
+ return status == GeoParseStatus::GEO_PARSE_OK ? std::move(shape) : nullptr;
+}
+
+static bool any_geo_argument_is_null(const Block& block, const ColumnNumbers&
arguments,
+ size_t row) {
+ return std::ranges::any_of(arguments, [&](auto argument) {
+ return block.get_by_position(argument).column->is_null_at(row);
+ });
+}
+
+static Status decode_calculation_shape(StringRef value, const DataTypePtr&
type,
+ std::unique_ptr<GeoShape>* shape,
+ const char* function_name) {
+ *shape = decode_geo_shape(value, type);
+ // POLYGON EMPTY decodes successfully, but distance helpers retain
infinity/max sentinels.
+ // Reject it for typed calculations just like empty points/lines, leaving
raw WKB untouched.
+ if (is_spatial_type(type) &&
+ (*shape == nullptr ||
+ ((*shape)->type() == GeoShapeType::GEO_SHAPE_POLYGON &&
+ assert_cast<const GeoPolygon&>(**shape).polygon()->num_loops() ==
0))) {
+ return Status::NotSupported("Function {} does not support this
Geography WKB shape",
+ function_name);
+ }
+ return Status::OK();
+}
+
+// Calculations use canonical positions; raw spatial accessors still return
the original WKB.
+static S2Point canonical_geography_point(const S2Point& point) {
+ const auto coordinate = S2LatLng(point);
+ const double latitude = coordinate.lat().radians();
+ // All longitudes at a pole and +/-180 at the date line denote the same
position.
+ if (std::abs(latitude) == M_PI_2) {
+ return {0, 0, latitude > 0 ? 1.0 : -1.0};
+ }
+ if (coordinate.lng().radians() == M_PI) {
+ // Flip only sin(longitude): rebuilding latitude can change its last
bit.
+ return {point.x(), -point.y(), point.z()};
+ }
+ return point;
+}
+
+static void geography_shape_index(const GeoShape& shape, MutableS2ShapeIndex*
index) {
+ switch (shape.type()) {
+ case GeoShapeType::GEO_SHAPE_LINE_STRING: {
+ const auto* line = assert_cast<const GeoLine&>(shape).polyline();
+ std::vector<S2Point> vertices;
+ for (int i = 0; i < line->num_vertices(); ++i) {
+ vertices.push_back(canonical_geography_point(line->vertex(i)));
+ }
+ index->Add(std::make_unique<S2LaxPolylineShape>(vertices));
+ break;
+ }
+ case GeoShapeType::GEO_SHAPE_POLYGON: {
+ const auto* polygon = assert_cast<const GeoPolygon&>(shape).polygon();
+ std::vector<S2LaxPolygonShape::Loop> loops;
+ for (int i = 0; i < polygon->num_loops(); ++i) {
+ auto& vertices = loops.emplace_back();
+ const auto* loop = polygon->loop(i);
+ for (int j = 0; j < loop->num_vertices(); ++j) {
+
vertices.push_back(canonical_geography_point(loop->oriented_vertex(j)));
+ }
+ }
+ index->Add(std::make_unique<S2LaxPolygonShape>(loops));
+ break;
+ }
+ default:
+ // decode_calculation_shape admits only nonempty
Point/LineString/Polygon WKB.
+ throw Exception(ErrorCode::INTERNAL_ERROR, "Unexpected Geography
calculation shape");
+ }
+}
+
+static bool geography_point_relation(std::string_view name, const GeoShape&
lhs,
+ const GeoShape& rhs) {
+ const auto& point_shape = lhs.type() == GeoShapeType::GEO_SHAPE_POINT ?
lhs : rhs;
+ const auto& other = lhs.type() == GeoShapeType::GEO_SHAPE_POINT ? rhs :
lhs;
+ const auto point =
+ canonical_geography_point(*assert_cast<const
GeoPoint&>(point_shape).point());
+ bool interior = false;
+ bool boundary = false;
+ if (other.type() == GeoShapeType::GEO_SHAPE_POINT) {
+ interior = point == canonical_geography_point(*assert_cast<const
GeoPoint&>(other).point());
+ } else if (other.type() == GeoShapeType::GEO_SHAPE_LINE_STRING) {
+ const auto* line = assert_cast<const GeoLine&>(other).polyline();
+ bool member = false;
+ // S2 Boolean operations symbolically move points off edge interiors.
Use its
+ // unperturbed distance predicate on every real geodesic edge instead.
+ for (int i = 0; i + 1 < line->num_vertices(); ++i) {
+ member = member ||
+ s2pred::CompareEdgeDistance(point,
canonical_geography_point(line->vertex(i)),
+
canonical_geography_point(line->vertex(i + 1)),
+ S1ChordAngle::Zero()) <= 0;
Review Comment:
Completed the near-antipodal follow-up in c7d31ed3989, retained in
af66f59ab50. Point membership and polygon boundaries now use
original-coordinate arc predicates. Added midpoint, nearby-distinct, endpoint
and both-order coverage; 115 geo/WKB UTs and both targeted suites passed.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
