Fix tensor.searchsorted for x1 with strides and scalar (0D) x2 (#1693)

* Adds code to handle edge case of strided input and scalar `needle` in `searchsorted.cpp`

* Adds a test for fix to gh-1689

Author: ndgrigorian

dpctl/tensor/libtensor/source/sorting/searchsorted.cpp

@@ -340,22 +340,29 @@ py_searchsorted(const dpctl::tensor::usm_ndarray &hay,
     int simplified_nd = needles_nd;
 
     using shT = std::vector<py::ssize_t>;
-
     shT simplified_common_shape;
     shT simplified_needles_strides;
     shT simplified_positions_strides;
     py::ssize_t needles_offset(0);
     py::ssize_t positions_offset(0);
 
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        // modified by refernce
-        simplified_nd,
-        // read-only inputs
-        needles_shape_ptr, needles_strides, positions_strides,
-        // output, modified by reference
-        simplified_common_shape, simplified_needles_strides,
-        simplified_positions_strides, needles_offset, positions_offset);
-
+    if (simplified_nd == 0) {
+        // needles and positions have same nd
+        simplified_nd = 1;
+        simplified_common_shape.push_back(1);
+        simplified_needles_strides.push_back(0);
+        simplified_positions_strides.push_back(0);
+    }
+    else {
+        dpctl::tensor::py_internal::simplify_iteration_space(
+            // modified by refernce
+            simplified_nd,
+            // read-only inputs
+            needles_shape_ptr, needles_strides, positions_strides,
+            // output, modified by reference
+            simplified_common_shape, simplified_needles_strides,
+            simplified_positions_strides, needles_offset, positions_offset);
+    }
     std::vector<sycl::event> host_task_events;
     host_task_events.reserve(2);
 

dpctl/tests/test_usm_ndarray_searchsorted.py

@@ -355,3 +355,19 @@ def test_out_of_bound_sorter_values():
     p = dpt.searchsorted(x, x2, sorter=sorter)
     # verify that they were applied with mode="wrap"
     assert dpt.all(p == dpt.arange(3, dtype=p.dtype))
+
+
+def test_searchsorted_strided_scalar_needle():
+    get_queue_or_skip()
+
+    a_max = 255
+
+    hay_stack = dpt.flip(
+        dpt.repeat(dpt.arange(a_max - 1, -1, -1, dtype=dpt.int32), 4)
+    )
+    needles_np = np.squeeze(
+        np.random.randint(0, a_max, dtype=dpt.int32, size=1), axis=0
+    )
+    needles = dpt.asarray(needles_np)
+
+    _check(hay_stack, needles, needles_np)