BUG: Spurious FutureWarning when using pd.read_json()

[krassowski]

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
s = '{"A":{"0":"X","Y":"Y"}}'
pd.read_json(io.StringIO(s), typ='frame', orient='records')

Issue Description

read_json() works correctly but generates three spurious and annoying warnings:

FutureWarning: The behavior of 'to_datetime' with 'unit' when parsing strings is deprecated. In a future version, strings will be parsed as datetime strings, matching the behavior without a 'unit'. To retain the old behavior, explicitly cast ints or floats to numeric type before calling to_datetime.
  pd.read_json(io.StringIO(s), typ='frame', orient='records')
FutureWarning: The behavior of 'to_datetime' with 'unit' when parsing strings is deprecated. In a future version, strings will be parsed as datetime strings, matching the behavior without a 'unit'. To retain the old behavior, explicitly cast ints or floats to numeric type before calling to_datetime.
  pd.read_json(io.StringIO(s), typ='frame', orient='records')
FutureWarning: The behavior of 'to_datetime' with 'unit' when parsing strings is deprecated. In a future version, strings will be parsed as datetime strings, matching the behavior without a 'unit'. To retain the old behavior, explicitly cast ints or floats to numeric type before calling to_datetime.
  pd.read_json(io.StringIO(s), typ='frame', orient='records')

These warnings are spurious because the usage of to_datetime is not controlled by user. The warnings are annoying because there are so many of them.

This was previously raised on Stack Overflow here 3 months ago and it accumulated almost 400 visits: https://stackoverflow.com/questions/78454457/pandas-read-json-future-warning-the-behavior-of-to-datetime-with-unit-when

Expected Behavior

The warnings are suppressed.

Installed Versions

INSTALLED VERSIONS

commit : d9cdd2e
python : 3.10.10.final.0
python-bits : 64
OS : Linux
OS-release : 6.5.0-44-generic
Version : #44-Ubuntu SMP PREEMPT_DYNAMIC Fri Jun 7 15:10:09 UTC 2024
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_GB.UTF-8
LOCALE : en_GB.UTF-8

pandas : 2.2.2
numpy : 2.0.1
pytz : 2024.1
dateutil : 2.9.0.post0
setuptools : 65.5.0
pip : 22.3.1
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.1.4
IPython : 8.26.0
pandas_datareader : None
adbc-driver-postgresql: None
adbc-driver-sqlite : None
bs4 : 4.12.3
bottleneck : None
dataframe-api-compat : None
fastparquet : None
fsspec : None
gcsfs : None
matplotlib : None
numba : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyreadstat : None
python-calamine : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
zstandard : None
tzdata : 2024.1
qtpy : None
pyqt5 : None

[sgysherry]

I test with latest dev version, seems the problem is solved. Will take a closer look

[rhshadrach]

Thanks for the report, confirmed on 2.2.x. At a quick glance, it looks like this line:

pandas/pandas/io/json/_json.py

Line 1148 in 0851ac3

convert_dates=True,

should be passing self.convert_dates instead of being hard coded to True.

Additionally, we could catch the warning here:

pandas/pandas/io/json/_json.py

Line 1360 in 795cce2

new_data = to_datetime(new_data, errors="raise", unit=date_unit)

Further investigations and PRs to fix are welcome!

[KevsterAmp]

Take

[KevsterAmp]

seems like self.convert_dates also returns True by default.

The convert_axes param of pd.read_json() is True for all orient values except "Table"
https://pandas.pydata.org/docs/dev/reference/api/pandas.read_json.html

I was trying to see the difference between the pd.read_json() 2.2.x branch with main, can't really find much except that the specific FutureWarning is now removed from the main branch.

@rhshadrach Any ideas for a fix? Thanks

[rhshadrach]

seems like self.convert_dates also returns True by default.
...
The convert_axes param of pd.read_json() is True for all orient values except "Table"

I may be misunderstanding, but I think you're talking about the default value of convert_dates? The user can supply convert_dates=False to read_json. If we didn't have it hardcoded to True but rather used the argument the user passes, then I think they could suppress this warning. Of course, users shouldn't have to pass convert_dates=False to suppress the warning, but I think it's a better option than being forced to capture it.

Somewhat separately, it seems to me we should be adhering to a user's request when they pass convert_dates=False. That is not currently the case because it is hardcoded to True.

[KevsterAmp]

Thanks for the clarification @rhshadrach. I misunderstood the code on my previous message

[asishm]

I don't think this OP issue reproduces on main (was presumably fixed by #59124 - haven't run a bisect). I think it's better to create a new issue for the comment (speaking from the perspective of a release-notes reading user)

Somewhat separately, it seems to me we should be adhering to a user's request when they pass convert_dates=False. That is not currently the case because it is hardcoded to True.

which I believe is what @KevsterAmp's PR aims to do.

[rhshadrach]

I don't think this OP issue reproduces on main

I believe the warning has been removed in preparation of pandas 3.0. The warning still exists on 2.2, and I think this is eligible for a backport. In fact, we would skip a PR into main because this is no longer an issue on main, so a PR straight into either 2.2.x or 2.3.x.