BUG: Enlarging DataFrame with 2d indexer converts ints to floats

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 numpy as np

ct_arr = np.array([[70, 150], [66, 81]])
counts = pd.DataFrame(
    data=ct_arr,
    columns=["Outstanding", "Not Outstanding"],
    index=["Bank", "Credit Union"],
)
ctot = counts.copy()
# everything above this line is integers

ctot.loc["Total", :] = ctot.sum(axis=0)
ctot.loc[:, "Total"] = ctot.sum(axis=1)
print(ctot)

Issue Description

I generate a dataframe from an np.array() of integers. I want to calculate all the totals: by row, by column, and the grand total.

The code works, but the output is the float type.

Expected Behavior

Since all inputs are integers and I only add the numbers, I expected the output dataframe to also be only integers.

Installed Versions

INSTALLED VERSIONS

commit : e8093ba372f9adfe79439d90fe74b0b5b6dea9d6 python : 3.9.9.final.0 python-bits : 64 OS : Windows OS-release : 10 Version : 10.0.19044 machine : AMD64 processor : Intel64 Family 6 Model 158 Stepping 10, GenuineIntel byteorder : little LC_ALL : None LANG : None LOCALE : English_United States.1252

pandas : 1.4.3 numpy : 1.23.0 pytz : 2021.3 dateutil : 2.8.2 setuptools : 62.6.0 pip : 22.1.2 Cython : 0.29.30 pytest : None hypothesis : None sphinx : None blosc : None feather : None xlsxwriter : None lxml.etree : 4.7.1 html5lib : None pymysql : None psycopg2 : None jinja2 : 3.1.2 IPython : 8.4.0 pandas_datareader: None bs4 : 4.11.1 bottleneck : None brotli : None fastparquet : 0.8.1 fsspec : 2021.11.1 gcsfs : None markupsafe : 2.1.1 matplotlib : 3.5.2 numba : None numexpr : None odfpy : None openpyxl : None pandas_gbq : None pyarrow : 8.0.0 pyreadstat : None pyxlsb : None s3fs : None scipy : 1.8.1 snappy : None sqlalchemy : None tables : None tabulate : 0.8.10 xarray : 0.20.2 xlrd : 2.0.1 xlwt : None zstandard : None

[simonjayhawkins]

@jreback - why should expanding an integer column with an integer value coerce to float?

In case the examples are not clear, does not coerce to float with 1d indexer

df = pd.DataFrame([70, 150], dtype="i8")
df.loc["Total"] = df.sum(axis=0)
print(df.dtypes)
# 0    int64
# dtype: object

does coerce to float with 2d indexer

df = pd.DataFrame([70, 150], dtype="i8")
df.loc["Total", :] = df.sum(axis=0)
print(df.dtypes)
# 0    float64
# dtype: object

and only happens with axis=0, i.e. adding rows.

looks like a bug to me. These should both produce the same output?

[simonjayhawkins]

Thanks @FlorinAndrei for the report.

Note that the row expansion with a 2d indexer for .loc is causing the cast (the first total) and in the code sample this results in the floats in the row totals (second total and added column) and the DataFrame copy is also not relevant.

# single column DataFrame

df = pd.DataFrame([70, 150], dtype="i8")
df.loc["Total"] = df.sum(axis=0)
print(df.dtypes)
# 0    int64
# dtype: object

df = pd.DataFrame([70, 150], dtype="i8")
df.loc["Total", :] = df.sum(axis=0)
print(df.dtypes)
# 0    float64
# dtype: object

df = pd.DataFrame([70, 150], dtype="i8")
df.loc[:, "Total"] = df.sum(axis=1)
print(df.dtypes)
# 0        int64
# Total    int64
# dtype: object


# # multi-column DataFrame

df = pd.DataFrame([[70, 150], [66, 81]], dtype="i8")
df.loc["Total"] = df.sum(axis=0)
print(df.dtypes)
# 0    int64
# 1    int64
# dtype: object

df = pd.DataFrame([[70, 150], [66, 81]], dtype="i8")
df.loc["Total", :] = df.sum(axis=0)
print(df.dtypes)
# 0    float64
# 1    float64
# dtype: object

df = pd.DataFrame([[70, 150], [66, 81]], dtype="i8")
df.loc[:, "Total"] = df.sum(axis=1)
print(df.dtypes)
# 0        int64
# 1        int64
# Total    int64
# dtype: object

The result dtype from df.sum() is correct, so the title of this issue is misleading and the MRE could be reduced to

df = pd.DataFrame([70, 150], dtype="i8")
df.loc["Total", :] = pd.Series([220], dtype="i8")
df.dtypes
# 0    float64
# dtype: object

so it appears to be an indexing issue i.e. setting values with .loc with row expansion. (might be worth looking for similar issues)

contributions and PRs to fix welcome