BUG: assert_series_equal not raising on unequal series?

[MarcoGorelli]

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

left = pd.Series([1577840521123000])
right = pd.Series([1577840521123543])
pd.testing.assert_series_equal(left, right)

Issue Description

The above doesn't raise

Expected Behavior

It should raise

Installed Versions

In [4]: pd.show_versions()
/home/marcogorelli/tmp/.venv/lib/python3.10/site-packages/_distutils_hack/init.py:33: UserWarning: Setuptools is replacing distutils.
warnings.warn("Setuptools is replacing distutils.")

INSTALLED VERSIONS

commit : 51f3d03
python : 3.10.12.final.0
python-bits : 64
OS : Linux
OS-release : 5.10.102.1-microsoft-standard-WSL2
Version : #1 SMP Wed Mar 2 00:30:59 UTC 2022
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_GB.UTF-8
LOCALE : en_GB.UTF-8

pandas : 2.2.0.dev0+541.g51f3d03087
numpy : 1.25.1
pytz : 2023.3
dateutil : 2.8.2
setuptools : 67.6.1
pip : 23.1.2
Cython : None
pytest : 7.3.1
hypothesis : 6.82.4
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.1.2
IPython : 8.16.1
pandas_datareader : None
bs4 : 4.12.2
bottleneck : None
dataframe-api-compat: None
fastparquet : 2023.7.0
fsspec : 2023.6.0
gcsfs : None
matplotlib : 3.7.1
numba : 0.58.1
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 12.0.1
pyreadstat : None
python-calamine : None
pyxlsb : None
s3fs : None
scipy : 1.11.0
sqlalchemy : 2.0.20
tables : None
tabulate : None
xarray : None
xlrd : None
zstandard : None
tzdata : 2023.3
qtpy : None
pyqt5 : None

[EricJeannotte]

By default, the parameter check_exact=False, which means that default relative(1e-5) and absolute tolerances(1e-8) are applied according to documentation. Considering that the numbers are very large in your example, I believe the output is correct.

[MarcoGorelli]

from today's call, this check_exact / rtol / atol should not take effect for ints https://docs.google.com/document/d/1tGbTiYORHiSPgVMXawiweGJlBw5dOkVJLY-licoBmBU/edit

[lithomas1]

Labelling as blocker for 2.2, not because it is a blocker, but because it looks pretty nasty and we should definitely fix this.

[parthi-siva]

take

[parthi-siva]

Hi @MarcoGorelli
can we add a check like this in pandas/_testing/asserters.py

    if left.dtype == np.int64 and right.dtype == np.int64:
        check_exact = True

after adding it's getting raised

In [1]: import pandas as pd

In [2]: left = pd.Series([1577840521123000])
   ...: right = pd.Series([1577840521123543])

In [3]: pd.testing.assert_series_equal(left, right)
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
Cell In[3], line 1
----> 1 pd.testing.assert_series_equal(left, right)

    [... skipping hidden 2 frame]

File ~/Documents/pandas/pandas/_testing/asserters.py:672, in assert_numpy_array_equal.<locals>._raise(left, right, err_msg)
    670     diff = diff * 100.0 / left.size
    671     msg = f"{obj} values are different ({np.round(diff, 5)} %)"
--> 672     raise_assert_detail(obj, msg, left, right, index_values=index_values)
    674 raise AssertionError(err_msg)

File ~/Documents/pandas/pandas/_testing/asserters.py:602, in raise_assert_detail(obj, message, left, right, diff, first_diff, index_values)
    599 if first_diff is not None:
    600     msg += f"\n{first_diff}"
--> 602 raise AssertionError(msg)

AssertionError: Series are different

Series values are different (100.0 %)
[index]: [0]
[left]:  [1577840521123000]
[right]: [1577840521123543]

[MarcoGorelli]

nice - it would need to be more generic than if left.dtype == np.int64 and right.dtype == np.int64:, I think there's a is_integer_dtype function (or something like that) you might be able to use

[parthi-siva]

Yeah right. Thanks @MarcoGorelli create a PR. Once pls review