add string-specific ObjectEngine subclass for pre-processing of input values

Author: jorisvandenbossche

pandas/_libs/index.pyi

@@ -54,6 +54,7 @@ class UInt16Engine(IndexEngine): ...
 class UInt8Engine(IndexEngine): ...
 class ObjectEngine(IndexEngine): ...
 class StringEngine(IndexEngine): ...
+class StringObjectEngine(ObjectEngine): ...
 class DatetimeEngine(Int64Engine): ...
 class TimedeltaEngine(DatetimeEngine): ...
 class PeriodEngine(Int64Engine): ...

pandas/_libs/index.pyx

@@ -557,6 +557,16 @@ cdef class StringEngine(IndexEngine):
             raise KeyError(val)
         return str(val)
 
+cdef class StringObjectEngine(ObjectEngine):
+
+    cdef _check_type(self, object val):
+        if isinstance(val, str):
+            return val
+        elif checknull(val):
+            return np.nan
+        else:
+            raise KeyError(val)
+
 
 cdef class DatetimeEngine(Int64Engine):
 

pandas/core/indexes/base.py

@@ -876,6 +876,8 @@ def _engine(
             # ndarray[Any, Any]]" has no attribute "_ndarray"  [union-attr]
             target_values = self._data._ndarray  # type: ignore[union-attr]
         # TODO re-enable StringEngine for string dtype
+        elif is_string_dtype(self.dtype) and not is_object_dtype(self.dtype):
+            return libindex.StringObjectEngine(target_values)
 
         # error: Argument 1 to "ExtensionEngine" has incompatible type
         # "ndarray[Any, Any]"; expected "ExtensionArray"