BUG make hashtable.unique support readonly arrays

This problem was brought up in
pandas-dev#18773 and effectively comes
down to how Cython deals with readonly arrays. While it would be ideal
for Cython to fix the underlying problem in the meantime we can rely on
this.

pandas/_libs/hashtable_class_helper.pxi.in

@@ -255,10 +255,56 @@ dtypes = [('Float64', 'float64', 'val != val', True),
           ('UInt64', 'uint64', 'False', False),
           ('Int64', 'int64', 'val == iNaT', False)]
 
+def get_dispatch(dtypes):
+  for (name, dtype, null_condition, float_group) in dtypes:
+    unique_template = """\
+      cdef:
+         Py_ssize_t i, n = len(values)
+         int ret = 0
+         {dtype}_t val
+         khiter_t k
+         bint seen_na = 0
+         {name}Vector uniques = {name}Vector()
+         {name}VectorData *ud
+
+      ud = uniques.data
+
+      with nogil:
+          for i in range(n):
+              val = values[i]
+              IF {float_group}:
+                if val == val:
+                    k = kh_get_{dtype}(self.table, val)
+                    if k == self.table.n_buckets:
+                        kh_put_{dtype}(self.table, val, &ret)
+                        if needs_resize(ud):
+                            with gil:
+                                uniques.resize()
+                        append_data_{dtype}(ud, val)
+                elif not seen_na:
+                    seen_na = 1
+                    if needs_resize(ud):
+                        with gil:
+                            uniques.resize()
+                    append_data_{dtype}(ud, NAN)
+              ELSE:
+                k = kh_get_{dtype}(self.table, val)
+                if k == self.table.n_buckets:
+                    kh_put_{dtype}(self.table, val, &ret)
+                    if needs_resize(ud):
+                        with gil:
+                            uniques.resize()
+                    append_data_{dtype}(ud, val)
+      return uniques.to_array()
+      """
+
+    unique_template = unique_template.format(name=name, dtype=dtype, null_condition=null_condition, float_group=float_group)
+
+    yield (name, dtype, null_condition, float_group, unique_template)
 }}
 
 
-{{for name, dtype, null_condition, float_group in dtypes}}
+{{for name, dtype, null_condition, float_group, unique_template in get_dispatch(dtypes)}}
 
 cdef class {{name}}HashTable(HashTable):
 
@@ -450,48 +496,15 @@ cdef class {{name}}HashTable(HashTable):
         return np.asarray(labels), arr_uniques
 
     @cython.boundscheck(False)
-    def unique(self, {{dtype}}_t[:] values):
-        cdef:
-            Py_ssize_t i, n = len(values)
-            int ret = 0
-            {{dtype}}_t val
-            khiter_t k
-            bint seen_na = 0
-            {{name}}Vector uniques = {{name}}Vector()
-            {{name}}VectorData *ud
+    def unique(self, ndarray[{{dtype}}_t, ndim=1] values):
+      if values.flags.writeable:
+        return self.unique_memview(values)
 
-        ud = uniques.data
+{{unique_template}}
 
-        with nogil:
-            for i in range(n):
-                val = values[i]
-
-                {{if float_group}}
-                if val == val:
-                    k = kh_get_{{dtype}}(self.table, val)
-                    if k == self.table.n_buckets:
-                        kh_put_{{dtype}}(self.table, val, &ret)
-                        if needs_resize(ud):
-                            with gil:
-                                uniques.resize()
-                        append_data_{{dtype}}(ud, val)
-                elif not seen_na:
-                    seen_na = 1
-                    if needs_resize(ud):
-                        with gil:
-                            uniques.resize()
-                    append_data_{{dtype}}(ud, NAN)
-                {{else}}
-                k = kh_get_{{dtype}}(self.table, val)
-                if k == self.table.n_buckets:
-                    kh_put_{{dtype}}(self.table, val, &ret)
-                    if needs_resize(ud):
-                        with gil:
-                            uniques.resize()
-                    append_data_{{dtype}}(ud, val)
-                {{endif}}
-
-        return uniques.to_array()
+    @cython.boundscheck(False)
+    def unique_memview(self, {{dtype}}_t[:] values):
+{{unique_template}}
 
 {{endfor}}
 

pandas/tests/reshape/test_tile.py

@@ -512,6 +512,16 @@ def f():
         tm.assert_numpy_array_equal(
             mask, np.array([False, True, True, True, True]))
 
+    def test_cut_read_only(self):
+        readonly = np.arange(0, 100, 10)
+        readonly.flags.writeable = False
+
+        mutable = np.arange(0, 100, 10)
+
+        one_to_hundred = np.arange(100)
+        tm.assert_categorical_equal(cut(one_to_hundred, readonly), \
+                                    cut(one_to_hundred, mutable))
+
 
 def curpath():
     pth, _ = os.path.split(os.path.abspath(__file__))