BUG: TypeError: Int64 when using pd.pivot_table with Int64 values

[mattharrison]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd
import io
data = ''',country_live,employment_status,age
26983,United States,Fully employed by a company / organization,21
51776,Turkey,Working student,18
53092,India,Fully employed by a company / organization,21
44612,France,Fully employed by a company / organization,21
7043,United States,"Self-employed (a person earning income directly from one's own business, trade, or profession)",60
50421,Other country,Fully employed by a company / organization,21
2552,United Kingdom,Fully employed by a company / organization,30
21166,China,Fully employed by a company / organization,21
29144,Italy,Fully employed by a company / organization,40
36828,United States,Fully employed by a company / organization,40'''

df = pd.read_csv(io.StringIO(data))
(df
 .astype({'age': 'Int64'})
 .pivot_table(index='country_live', columns='employment_status',
     values='age', aggfunc='mean')
)

Issue Description

This aggregation raises a TypeError: Int64

However, it you change age to int64 the code works. Also note that it works if you perform the same operation via .groupby/.unstack:

ie:

(df
 #.astype({'age': 'Int64'})
 .pivot_table(index='country_live', columns='employment_status',
     values='age', aggfunc='mean')
)

This also works:

(df
 .astype({'age': 'Int64'})
 .groupby(['country_live', 'employment_status'])
 .age
 .mean()
 .unstack()
)

Expected Behavior

I expect to see the following result:

country_live	Fully employed by a company / organization	Self-employed (a person earning income directly from one's own business, trade, or profession)	Working student
China	21.0
France	21.0
India	21.0
Italy	40.0
Other country	21.0
Turkey			18.0
United Kingdom	30.0
United States	30.5	60.0
​

Installed Versions

INSTALLED VERSIONS

commit : 4bfe3d0
python : 3.8.10.final.0
python-bits : 64
OS : Linux
OS-release : 5.10.16.3-microsoft-standard-WSL2
Version : #1 SMP Fri Apr 2 22:23:49 UTC 2021
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : C.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.4.2
numpy : 1.19.4
pytz : 2022.1
dateutil : 2.8.1
pip : 20.0.2
setuptools : 44.0.0
Cython : 0.29.24
pytest : 7.1.1
hypothesis : 6.8.1
sphinx : 4.3.2
blosc : None
feather : None
xlsxwriter : 1.3.7
lxml.etree : 4.6.2
html5lib : 1.1
pymysql : None
psycopg2 : 2.9.3
jinja2 : 3.0.3
IPython : 7.19.0
pandas_datareader: None
bs4 : 4.10.0
bottleneck : None
brotli : None
fastparquet : None
fsspec : 2021.07.0
gcsfs : None
markupsafe : 2.1.1
matplotlib : 3.3.3
numba : 0.52.0
numexpr : 2.8.1
odfpy : None
openpyxl : 3.0.5
pandas_gbq : None
pyarrow : 8.0.0
pyreadstat : 1.1.4
pyxlsb : None
s3fs : None
scipy : 1.5.4
snappy : None
sqlalchemy : 1.4.27
tables : None
tabulate : 0.8.9
xarray : 0.18.2
xlrd : 2.0.1
xlwt : None
zstandard : None

[simonjayhawkins]

Thanks @mattharrison for the report.

Note: this was giving the expected result in 1.2.5 with Float64 dtype for mean ages (np.float64 in 1.1.5 and before)

[simonjayhawkins]

this was giving the expected result in 1.2.5

first bad commit: [4340689] REF: enforce annotation in maybe_downcast_to_dtype (#40982)

cc @jbrockmendel

[Mehgarg]

take