ENH: Compile Factorizer class for all numeric dtypes (pandas-dev#49624)

* ENH: Compile Factorizer class for all numeric dtypes

* Fix test

* Fix test

* Add factorize to base class

* Remove ignores

* Move factorizer

Author: phofl

pandas/_libs/hashtable.pyi

@@ -15,29 +15,65 @@ class Factorizer:
     count: int
     def __init__(self, size_hint: int) -> None: ...
     def get_count(self) -> int: ...
-
-class ObjectFactorizer(Factorizer):
-    table: PyObjectHashTable
-    uniques: ObjectVector
     def factorize(
         self,
-        values: npt.NDArray[np.object_],
+        values: np.ndarray,
         sort: bool = ...,
         na_sentinel=...,
         na_value=...,
     ) -> npt.NDArray[np.intp]: ...
 
+class ObjectFactorizer(Factorizer):
+    table: PyObjectHashTable
+    uniques: ObjectVector
+
 class Int64Factorizer(Factorizer):
     table: Int64HashTable
     uniques: Int64Vector
-    def factorize(
-        self,
-        values: np.ndarray,  # const int64_t[:]
-        sort: bool = ...,
-        na_sentinel=...,
-        na_value=...,
-        mask=...,
-    ) -> npt.NDArray[np.intp]: ...
+
+class UInt64Factorizer(Factorizer):
+    table: UInt64HashTable
+    uniques: UInt64Vector
+
+class Int32Factorizer(Factorizer):
+    table: Int32HashTable
+    uniques: Int32Vector
+
+class UInt32Factorizer(Factorizer):
+    table: UInt32HashTable
+    uniques: UInt32Vector
+
+class Int16Factorizer(Factorizer):
+    table: Int16HashTable
+    uniques: Int16Vector
+
+class UInt16Factorizer(Factorizer):
+    table: UInt16HashTable
+    uniques: UInt16Vector
+
+class Int8Factorizer(Factorizer):
+    table: Int8HashTable
+    uniques: Int8Vector
+
+class UInt8Factorizer(Factorizer):
+    table: UInt8HashTable
+    uniques: UInt8Vector
+
+class Float64Factorizer(Factorizer):
+    table: Float64HashTable
+    uniques: Float64Vector
+
+class Float32Factorizer(Factorizer):
+    table: Float32HashTable
+    uniques: Float32Vector
+
+class Complex64Factorizer(Factorizer):
+    table: Complex64HashTable
+    uniques: Complex64Vector
+
+class Complex128Factorizer(Factorizer):
+    table: Complex128HashTable
+    uniques: Complex128Vector
 
 class Int64Vector:
     def __init__(self, *args) -> None: ...

pandas/_libs/hashtable.pyx

@@ -79,6 +79,9 @@ cdef class Factorizer:
     def get_count(self) -> int:
         return self.count
 
+    def factorize(self, values, na_sentinel=-1, na_value=None, mask=None) -> np.ndarray:
+        raise NotImplementedError
+
 
 cdef class ObjectFactorizer(Factorizer):
     cdef public:
@@ -90,7 +93,7 @@ cdef class ObjectFactorizer(Factorizer):
         self.uniques = ObjectVector()
 
     def factorize(
-        self, ndarray[object] values, na_sentinel=-1, na_value=None
+        self, ndarray[object] values, na_sentinel=-1, na_value=None, mask=None
     ) -> np.ndarray:
         """
 
@@ -109,6 +112,9 @@ cdef class ObjectFactorizer(Factorizer):
         cdef:
             ndarray[intp_t] labels
 
+        if mask is not None:
+            raise NotImplementedError("mask not supported for ObjectFactorizer.")
+
         if self.uniques.external_view_exists:
             uniques = ObjectVector()
             uniques.extend(self.uniques.to_array())
@@ -117,41 +123,3 @@ cdef class ObjectFactorizer(Factorizer):
                                        self.count, na_sentinel, na_value)
         self.count = len(self.uniques)
         return labels
-
-
-cdef class Int64Factorizer(Factorizer):
-    cdef public:
-        Int64HashTable table
-        Int64Vector uniques
-
-    def __cinit__(self, size_hint: int):
-        self.table = Int64HashTable(size_hint)
-        self.uniques = Int64Vector()
-
-    def factorize(self, const int64_t[:] values,
-                  na_sentinel=-1, na_value=None, object mask=None) -> np.ndarray:
-        """
-        Returns
-        -------
-        ndarray[intp_t]
-
-        Examples
-        --------
-        Factorize values with nans replaced by na_sentinel
-
-        >>> fac = Int64Factorizer(3)
-        >>> fac.factorize(np.array([1,2,3]), na_sentinel=20)
-        array([0, 1, 2])
-        """
-        cdef:
-            ndarray[intp_t] labels
-
-        if self.uniques.external_view_exists:
-            uniques = Int64Vector()
-            uniques.extend(self.uniques.to_array())
-            self.uniques = uniques
-        labels = self.table.get_labels(values, self.uniques,
-                                       self.count, na_sentinel,
-                                       na_value=na_value, mask=mask)
-        self.count = len(self.uniques)
-        return labels

pandas/_libs/hashtable_class_helper.pxi.in

@@ -101,6 +101,7 @@ from pandas._libs.khash cimport (
 from pandas._libs.tslibs.util cimport get_c_string
 from pandas._libs.missing cimport C_NA
 
+
 {{py:
 
 # name, dtype, c_type
@@ -876,6 +877,44 @@ cdef class {{name}}HashTable(HashTable):
         return np.asarray(labels), arr_uniques
     {{endif}}
 
+
+cdef class {{name}}Factorizer(Factorizer):
+    cdef public:
+        {{name}}HashTable table
+        {{name}}Vector uniques
+
+    def __cinit__(self, size_hint: int):
+        self.table = {{name}}HashTable(size_hint)
+        self.uniques = {{name}}Vector()
+
+    def factorize(self, const {{c_type}}[:] values,
+                  na_sentinel=-1, na_value=None, object mask=None) -> np.ndarray:
+        """
+        Returns
+        -------
+        ndarray[intp_t]
+
+        Examples
+        --------
+        Factorize values with nans replaced by na_sentinel
+
+        >>> fac = {{name}}Factorizer(3)
+        >>> fac.factorize(np.array([1,2,3], dtype="{{dtype}}"), na_sentinel=20)
+        array([0, 1, 2])
+        """
+        cdef:
+            ndarray[intp_t] labels
+
+        if self.uniques.external_view_exists:
+            uniques = {{name}}Vector()
+            uniques.extend(self.uniques.to_array())
+            self.uniques = uniques
+        labels = self.table.get_labels(values, self.uniques,
+                                       self.count, na_sentinel,
+                                       na_value=na_value, mask=mask)
+        self.count = len(self.uniques)
+        return labels
+
 {{endfor}}
 
 

pandas/core/reshape/merge.py

@@ -2362,14 +2362,8 @@ def _factorize_keys(
 
     rizer = klass(max(len(lk), len(rk)))
 
-    # Argument 1 to "factorize" of "ObjectFactorizer" has incompatible type
-    # "Union[ndarray[Any, dtype[signedinteger[_64Bit]]],
-    # ndarray[Any, dtype[object_]]]"; expected "ndarray[Any, dtype[object_]]"
-    llab = rizer.factorize(lk)  # type: ignore[arg-type]
-    # Argument 1 to "factorize" of "ObjectFactorizer" has incompatible type
-    # "Union[ndarray[Any, dtype[signedinteger[_64Bit]]],
-    # ndarray[Any, dtype[object_]]]"; expected "ndarray[Any, dtype[object_]]"
-    rlab = rizer.factorize(rk)  # type: ignore[arg-type]
+    llab = rizer.factorize(lk)
+    rlab = rizer.factorize(rk)
     assert llab.dtype == np.dtype(np.intp), llab.dtype
     assert rlab.dtype == np.dtype(np.intp), rlab.dtype