Use infer_dtype to detect strings; add new regex for tests

Author: kinshukdua

pandas/core/nanops.py

@@ -20,6 +20,7 @@
     iNaT,
     lib,
 )
+from pandas._libs.lib import infer_dtype
 from pandas._typing import (
     ArrayLike,
     Dtype,
@@ -43,7 +44,6 @@
     is_numeric_dtype,
     is_object_dtype,
     is_scalar,
-    is_string_dtype,
     is_timedelta64_dtype,
     needs_i8_conversion,
     pandas_dtype,
@@ -693,10 +693,9 @@ def nanmean(
 
     count = _get_counts(values.shape, mask, axis, dtype=dtype_count)
     the_sum = values.sum(axis, dtype=dtype_sum)
-    if isinstance(the_sum, str) or is_string_dtype(the_sum):
+    if infer_dtype(the_sum) in ("string", "byte", "mixed-integer", "mixed"):
         raise TypeError("cannot find the mean of type 'str'")
-    else:
-        _ensure_numeric(the_sum)
+    the_sum = _ensure_numeric(the_sum)
 
     if axis is not None and getattr(the_sum, "ndim", False):
         count = cast(np.ndarray, count)

pandas/tests/apply/test_invalid_arg.py

@@ -257,7 +257,10 @@ def test_agg_cython_table_raises_frame(df, func, expected, axis):
 )
 def test_agg_cython_table_raises_series(series, func, expected):
     # GH21224
-    msg = r"[Cc]ould not convert|can't multiply sequence by non-int of type"
+    msg = (
+        r"[Cc]ould not convert|can't multiply sequence by non-int of type"
+        r"|cannot find the mean of type 'str'"
+    )
     with pytest.raises(expected, match=msg):
         # e.g. Series('a b'.split()).cumprod() will raise
         series.agg(func)

pandas/tests/groupby/aggregate/test_cython.py

@@ -92,7 +92,7 @@ def test_cython_agg_nothing_to_agg():
     with pytest.raises(NotImplementedError, match="does not implement"):
         frame.groupby("a")["b"].mean(numeric_only=True)
 
-    with pytest.raises(TypeError, match="Could not convert (foo|bar)*"):
+    with pytest.raises(TypeError, match="cannot find the mean of*"):
         frame.groupby("a")["b"].mean()
 
     frame = DataFrame({"a": np.random.randint(0, 5, 50), "b": ["foo", "bar"] * 25})