[SEDONA-508] Add ST_CrossesDateLine (#1259)

* Implement ST_CrossesDateLine

* Add docs

* Refactor crossesDateLine(); Fix python test

* Port ST_CrossesDateLine from Predicates to Functions

* Fix typo

common/src/main/java/org/apache/sedona/common/Functions.java

@@ -218,6 +218,51 @@ public static int bestSRID(Geometry geometry) {
         return Spheroid.EPSG_WORLD_MERCATOR;
     }
 
+    /**
+     * Checks if a geometry crosses the International Date Line.
+     *
+     * @param geometry The geometry to check.
+     * @return True if the geometry crosses the Date Line, false otherwise.
+     */
+    public static boolean crossesDateLine(Geometry geometry) {
+        if (geometry == null || geometry.isEmpty()) {
+            return false;
+        }
+
+        CoordinateSequenceFilter filter = new CoordinateSequenceFilter() {
+            private Coordinate previous = null;
+            private boolean crossesDateLine = false;
+
+            @Override
+            public void filter(CoordinateSequence seq, int i) {
+                if (i == 0) {
+                    previous = seq.getCoordinateCopy(i);
+                    return;
+                }
+
+                Coordinate current = seq.getCoordinateCopy(i);
+                if (Math.abs(current.x - previous.x) > 180) {
+                    crossesDateLine = true;
+                }
+
+                previous = current;
+            }
+
+            @Override
+            public boolean isDone() {
+                return crossesDateLine;
+            }
+
+            @Override
+            public boolean isGeometryChanged() {
+                return false;
+            }
+        };
+
+        geometry.apply(filter);
+        return filter.isDone();
+    }
+
     public static Geometry envelope(Geometry geometry) {
         return geometry.getEnvelope();
     }

common/src/main/java/org/apache/sedona/common/Predicates.java

@@ -13,9 +13,11 @@
  */
 package org.apache.sedona.common;
 
-import org.locationtech.jts.geom.Geometry;
+import org.locationtech.jts.geom.*;
 import org.apache.sedona.common.sphere.Spheroid;
 
+import java.util.concurrent.atomic.AtomicBoolean;
+
 public class Predicates {
     public static boolean contains(Geometry leftGeometry, Geometry rightGeometry) {
         return leftGeometry.contains(rightGeometry);
@@ -62,4 +64,5 @@ public static boolean dWithin(Geometry leftGeometry, Geometry rightGeometry, dou
             return leftGeometry.isWithinDistance(rightGeometry, distance);
         }
     }
+
 }

common/src/test/java/org/apache/sedona/common/PredicatesTest.java

@@ -22,7 +22,11 @@
 import org.locationtech.jts.geom.Coordinate;
 import org.locationtech.jts.geom.Geometry;
 import org.locationtech.jts.geom.GeometryFactory;
+import org.locationtech.jts.io.ParseException;
 
+import static org.apache.sedona.common.Constructors.geomFromEWKT;
+import static org.apache.sedona.common.Constructors.geomFromWKT;
+import static org.apache.sedona.common.Functions.crossesDateLine;
 import static org.junit.Assert.*;
 
 public class PredicatesTest extends TestBase {
@@ -72,5 +76,33 @@ public void testDWithinSpheroid() {
         assertTrue(actual);
     }
 
+    @Test
+    public void testCrossesDateLine() throws ParseException {
+        Geometry geom1 = geomFromEWKT("LINESTRING(170 30, -170 30)");
+        Geometry geom2 = geomFromEWKT("LINESTRING(-120 30, -130 40)");
+        Geometry geom3 = geomFromEWKT("POLYGON((175 10, -175 10, -175 -10, 175 -10, 175 10))");
+        Geometry geom4 = geomFromEWKT("POLYGON((-120 10, -130 10, -130 -10, -120 -10, -120 10))");
+        Geometry geom5 = geomFromEWKT("POINT(180 30)");
+        Geometry geom6 = geomFromEWKT("POLYGON((160 20, 180 20, 180 -20, 160 -20, 160 20), (165 15, 175 15, 175 -15, 165 -15, 165 15))");
+        Geometry geom8 = geomFromEWKT("POLYGON((170 -10, -170 -10, -170 10, 170 10, 170 -10), (175 -5, -175 -5, -175 5, 175 5, 175 -5))");
+
+        // Multi-geometry test cases
+        Geometry multiGeom1 = geomFromEWKT("MULTILINESTRING((170 30, -170 30), (-120 30, -130 40))");
+        Geometry multiGeom2 = geomFromEWKT("MULTIPOLYGON(((175 10, -175 10, -175 -10, 175 -10, 175 10)), ((-120 10, -130 10, -130 -10, -120 -10, -120 10)))");
+        Geometry multiGeom3 = geomFromEWKT("MULTIPOINT((180 30), (170 -20))");
+        Geometry multiGeom4 = geomFromEWKT("MULTIPOLYGON(((160 20, 180 20, 180 -20, 160 -20, 160 20)), ((-120 10, -130 10, -130 -10, -120 -10, -120 10)))");
+
+        assertEquals(true, crossesDateLine(geom1));
+        assertEquals(false, crossesDateLine(geom2));
+        assertEquals(true, crossesDateLine(geom3));
+        assertEquals(false, crossesDateLine(geom4));
+        assertEquals(false, crossesDateLine(geom5));
+        assertEquals(false, crossesDateLine(geom6));
+        assertEquals(true, crossesDateLine(geom8));
+        assertEquals(true, crossesDateLine(multiGeom1));
+        assertEquals(true, crossesDateLine(multiGeom2));
+        assertEquals(false, crossesDateLine(multiGeom3));
+        assertEquals(false, crossesDateLine(multiGeom4));
+    }
 
 }

docs/api/flink/Function.md

@@ -835,6 +835,35 @@ Output:
 2
 ```
 
+## ST_CrossesDateLine
+
+Introduction: This function determines if a given geometry crosses the International Date Line. It operates by checking if the difference in longitude between any pair of consecutive points in the geometry exceeds 180 degrees. If such a difference is found, it is assumed that the geometry crosses the Date Line. It returns true if the geometry crosses the Date Line, and false otherwise.
+
+!!!note
+    The function assumes that the provided geometry is in lon/lat coordinate reference system where longitude values range from -180 to 180 degrees.
+
+!!!note
+    For multi-geometries (e.g., MultiPolygon, MultiLineString), this function will return true if any one of the geometries within the multi-geometry crosses the International Date Line.
+
+Format: `ST_CrossesDateLine(geometry: Geometry)`
+
+Since: `v1.6.0`
+
+SQL Example:
+
+```sql
+SELECT ST_CrossesDateLine(ST_GeomFromWKT('LINESTRING(170 30, -170 30)'))
+```
+
+Output:
+
+```sql
+true
+```
+
+!!!Warning
+    For geometries that span more than 180 degrees in longitude without actually crossing the Date Line, this function may still return true, indicating a crossing.
+
 ## ST_Dimension
 
 Introduction: Return the topological dimension of this Geometry object, which must be less than or equal to the coordinate dimension. OGC SPEC s2.1.1.1 - returns 0 for POINT, 1 for LINESTRING, 2 for POLYGON, and the largest dimension of the components of a GEOMETRYCOLLECTION. If the dimension is unknown (e.g. for an empty GEOMETRYCOLLECTION) 0 is returned.

docs/api/snowflake/vector-data/Function.md

@@ -600,6 +600,35 @@ Output:
 2
 ```
 
+## ST_CrossesDateLine
+
+Introduction: This function determines if a given geometry crosses the International Date Line. It operates by checking if the difference in longitude between any pair of consecutive points in the geometry exceeds 180 degrees. If such a difference is found, it is assumed that the geometry crosses the Date Line. It returns true if the geometry crosses the Date Line, and false otherwise.
+
+!!!note
+    The function assumes that the provided geometry is in lon/lat coordinate reference system where longitude values range from -180 to 180 degrees.
+
+!!!note
+    For multi-geometries (e.g., MultiPolygon, MultiLineString), this function will return true if any one of the geometries within the multi-geometry crosses the International Date Line.
+
+Format: `ST_CrossesDateLine(geometry: Geometry)`
+
+Since: `v1.6.0`
+
+SQL Example:
+
+```sql
+SELECT ST_CrossesDateLine(ST_GeomFromWKT('LINESTRING(170 30, -170 30)'))
+```
+
+Output:
+
+```sql
+true
+```
+
+!!!Warning
+    For geometries that span more than 180 degrees in longitude without actually crossing the Date Line, this function may still return true, indicating a crossing.
+
 ## ST_Degrees
 
 Introduction: Convert an angle in radian to degrees.

docs/api/sql/Function.md

@@ -842,6 +842,35 @@ Output:
 2
 ```
 
+## ST_CrossesDateLine
+
+Introduction: This function determines if a given geometry crosses the International Date Line. It operates by checking if the difference in longitude between any pair of consecutive points in the geometry exceeds 180 degrees. If such a difference is found, it is assumed that the geometry crosses the Date Line. It returns true if the geometry crosses the Date Line, and false otherwise.
+
+!!!note
+    The function assumes that the provided geometry is in lon/lat coordinate reference system where longitude values range from -180 to 180 degrees.
+
+!!!note
+    For multi-geometries (e.g., MultiPolygon, MultiLineString), this function will return true if any one of the geometries within the multi-geometry crosses the International Date Line.
+
+Format: `ST_CrossesDateLine(geometry: Geometry)`
+
+Since: `v1.6.0`
+
+SQL Example:
+
+```sql
+SELECT ST_CrossesDateLine(ST_GeomFromWKT('LINESTRING(170 30, -170 30)'))
+```
+
+Output:
+
+```sql
+true
+```
+
+!!!Warning
+    For geometries that span more than 180 degrees in longitude without actually crossing the Date Line, this function may still return true, indicating a crossing.
+
 ## ST_Degrees
 
 Introduction: Convert an angle in radian to degrees.

flink/src/main/java/org/apache/sedona/flink/Catalog.java

@@ -53,6 +53,7 @@ public static UserDefinedFunction[] getFuncs() {
                 new Functions.ST_CollectionExtract(),
                 new Functions.ST_ConcaveHull(),
                 new Functions.ST_ConvexHull(),
+                new Functions.ST_CrossesDateLine(),
                 new Functions.ST_Envelope(),
                 new Functions.ST_Difference(),
                 new Functions.ST_Dimension(),

flink/src/main/java/org/apache/sedona/flink/expressions/Functions.java

@@ -172,6 +172,21 @@ public Geometry eval(@DataTypeHint(value = "RAW", bridgedTo = org.locationtech.j
         }
     }
 
+    public static class ST_CrossesDateLine extends ScalarFunction
+    {
+        /**
+         * Constructor for relation checking without duplicate removal
+         */
+        public ST_CrossesDateLine() {}
+
+        @DataTypeHint("Boolean")
+        public Boolean eval(@DataTypeHint(value = "RAW", bridgedTo = org.locationtech.jts.geom.Geometry.class) Object o)
+        {
+            Geometry geom = (Geometry) o;
+            return org.apache.sedona.common.Functions.crossesDateLine(geom);
+        }
+    }
+
     public static class ST_Envelope extends ScalarFunction {
         @DataTypeHint(value = "RAW", bridgedTo = org.locationtech.jts.geom.Geometry.class)
         public Geometry eval(@DataTypeHint(value = "RAW", bridgedTo = org.locationtech.jts.geom.Geometry.class) Object o) {

flink/src/test/java/org/apache/sedona/flink/FunctionTest.java

@@ -199,6 +199,33 @@ public void testConvexHull() {
         assertEquals("POLYGON ((0 0, 1 2, 3 2, 5 0, 0 0))", result.toString());
     }
 
+    @Test
+    public void testCrossesDateLine() {
+        // Test line crossing the Date Line
+        Table table1 = tableEnv.sqlQuery("SELECT ST_GeomFromWKT('LINESTRING(170 30, -170 30)') AS geom");
+        table1 = table1.select(call("ST_CrossesDateLine", $("geom")));
+        Boolean actual1 = (Boolean) first(table1).getField(0);
+        assertEquals(true, actual1);
+
+        // Test line not crossing the Date Line
+        Table table2 = tableEnv.sqlQuery("SELECT ST_GeomFromWKT('LINESTRING(-120 30, -130 40)') AS geom");
+        table2 = table2.select(call("ST_CrossesDateLine", $("geom")));
+        Boolean actual2 = (Boolean) first(table2).getField(0);
+        assertEquals(false, actual2);
+
+        // Test polygon crossing the Date Line
+        Table table3 = tableEnv.sqlQuery("SELECT ST_GeomFromWKT('POLYGON((175 10, -175 10, -175 -10, 175 -10, 175 10))') AS geom");
+        table3 = table3.select(call("ST_CrossesDateLine", $("geom")));
+        Boolean actual3 = (Boolean) first(table3).getField(0);
+        assertEquals(true, actual3);
+
+        // Test polygon not crossing the Date Line
+        Table table4 = tableEnv.sqlQuery("SELECT ST_GeomFromWKT('POLYGON((-120 10, -130 10, -130 -10, -120 -10, -120 10))') AS geom");
+        table4 = table4.select(call("ST_CrossesDateLine", $("geom")));
+        Boolean actual4 = (Boolean) first(table4).getField(0);
+        assertEquals(false, actual4);
+    }
+
     @Test
     public void testDifference() {
         Table lineTable = tableEnv.sqlQuery("SELECT ST_GeomFromWKT('LINESTRING(50 100, 50 200)') AS g1, ST_GeomFromWKT('LINESTRING(50 50, 50 150)') as g2");

python/sedona/sql/st_functions.py

@@ -47,6 +47,7 @@
     "ST_ClosestPoint",
     "ST_ConcaveHull",
     "ST_ConvexHull",
+    "ST_CrossesDateLine",
     "ST_Difference",
     "ST_Dimension",
     "ST_Distance",
@@ -457,6 +458,17 @@ def ST_ConvexHull(geometry: ColumnOrName) -> Column:
     """
     return _call_st_function("ST_ConvexHull", geometry)
 
+@validate_argument_types
+def ST_CrossesDateLine(a: ColumnOrName) -> Column:
+    """Check whether geometry a crosses the International Date Line.
+
+    :param a: Geometry to check crossing with.
+    :type a: ColumnOrName
+    :return: True if geometry a cross the dateline.
+    :rtype: Column
+    """
+    return _call_st_function("ST_CrossesDateLine", (a))
+
 @validate_argument_types
 def ST_Dimension(geometry: ColumnOrName):
     """Calculate the inherent dimension of a geometry column.

python/sedona/sql/st_predicates.py

@@ -67,7 +67,6 @@ def ST_Crosses(a: ColumnOrName, b: ColumnOrName) -> Column:
     """
     return _call_predicate_function("ST_Crosses", (a, b))
 
-
 @validate_argument_types
 def ST_Disjoint(a: ColumnOrName, b: ColumnOrName) -> Column:
     """Check whether two geometries are disjoint.

python/tests/sql/test_dataframe_api.py

@@ -86,6 +86,7 @@
     (stf.ST_ConcaveHull, ("geom", 1.0, True), "triangle_geom", "", "POLYGON ((1 1, 1 0, 0 0, 1 1))"),
     (stf.ST_ConvexHull, ("geom",), "triangle_geom", "", "POLYGON ((0 0, 1 1, 1 0, 0 0))"),
     (stf.ST_CoordDim, ("point",), "point_geom", "", 2),
+    (stf.ST_CrossesDateLine, ("line",), "line_crossing_dateline", "", True),
     (stf.ST_Difference, ("a", "b"), "overlapping_polys", "", "POLYGON ((1 0, 0 0, 0 1, 1 1, 1 0))"),
     (stf.ST_Dimension, ("geom",), "geometry_geom_collection", "", 1),
     (stf.ST_Distance, ("a", "b"), "two_points", "", 3.0),
@@ -241,6 +242,7 @@
     (stf.ST_CollectionExtract, (None,)),
     (stf.ST_ConcaveHull, (None, 1.0)),
     (stf.ST_ConvexHull, (None,)),
+    (stf.ST_CrossesDateLine, (None,)),
     (stf.ST_Difference, (None, "b")),
     (stf.ST_Difference, ("", None)),
     (stf.ST_Distance, (None, "")),
@@ -445,6 +447,8 @@ def base_df(self, request):
             return TestDataFrameAPI.spark.sql("SELECT ST_GeomFromWKT('POINT (0 0)') AS origin, ST_GeomFromWKT('POINT (1 0)') as point")
         elif request.param == "ny_seattle":
             return TestDataFrameAPI.spark.sql("SELECT ST_GeomFromWKT('POINT (-122.335167 47.608013)') AS seattle, ST_GeomFromWKT('POINT (-73.935242 40.730610)') as ny")
+        elif request.param == "line_crossing_dateline":
+            return TestDataFrameAPI.spark.sql("SELECT ST_GeomFromWKT('LINESTRING (179.95 30, -179.95 30)') AS line")
         raise ValueError(f"Invalid base_df name passed: {request.param}")
 
     def _id_test_configuration(val):

python/tests/sql/test_function.py

@@ -159,6 +159,20 @@ def test_st_centroid(self):
         function_df = self.spark.sql("select ST_Centroid(polygondf.countyshape) from polygondf")
         function_df.show()
 
+    def test_st_crossesdateline(self):
+        crosses_test_table = self.spark.sql(
+            "select ST_GeomFromWKT('POLYGON((175 10, -175 10, -175 -10, 175 -10, 175 10))') as geom")
+        crosses_test_table.createOrReplaceTempView("crossesTesttable")
+        crosses = self.spark.sql("select(ST_CrossesDateLine(geom)) from crossesTesttable")
+
+        not_crosses_test_table = self.spark.sql(
+            "select ST_GeomFromWKT('POLYGON((1 1, 4 1, 4 4, 1 4, 1 1))') as geom")
+        not_crosses_test_table.createOrReplaceTempView("notCrossesTesttable")
+        not_crosses = self.spark.sql("select(ST_CrossesDateLine(geom)) from notCrossesTesttable")
+
+        assert crosses.take(1)[0][0]
+        assert not not_crosses.take(1)[0][0]
+
     def test_st_length(self):
         polygon_wkt_df = self.spark.read.format("csv"). \
             option("delimiter", "\t"). \

snowflake-tester/src/test/java/org/apache/sedona/snowflake/snowsql/TestFunctions.java

@@ -271,6 +271,19 @@ public void test_ST_ConvexHull() {
                 "POLYGON ((10 8, 20 30, 100 190, 150 10, 10 8))"
         );
     }
+
+    @Test
+    public void test_ST_CrossesDateLine() {
+        registerUDF("ST_CrossesDateLine", byte[].class);
+        verifySqlSingleRes(
+                "SELECT SEDONA.ST_CrossesDateLine(sedona.ST_GeomFromWKT('POLYGON((175 10, -175 10, -175 -10, 175 -10, 175 10))'))",
+                true
+        );
+        verifySqlSingleRes(
+                "SELECT SEDONA.ST_CrossesDateLine(sedona.ST_GeomFromWKT('POLYGON((0 0, 0 1, 1 1, 1 0, 0 0))'))",
+                false
+        );
+    }
     @Test
     public void test_ST_Difference() {
         registerUDF("ST_Difference", byte[].class, byte[].class);