BUG: pd.Timestamp becomes int in .values printout

[aminnj]

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

print(pd.DataFrame([dict(
    a=1,
    b=pd.Timestamp('2023-01-01 12:00:00'),
    c=pd.Timestamp('2023-01-01 12:00:00'),
)]).values.tolist())
# [[1, Timestamp('2023-01-01 12:00:00'), Timestamp('2023-01-01 12:00:00')]]

print(pd.DataFrame([dict(
    b=pd.Timestamp('2023-01-01 12:00:00'),
    c=pd.Timestamp('2023-01-01 12:00:00'),
)]).values.tolist())
# [[1672574400000000000, 1672574400000000000]]

Issue Description

When all columns in a dataframe are pd.Timestamp, accessing values as strings yields integers. However, when a non-pd.Timestamp column exists in the dataframe, the string representation is as expected.

Expected Behavior

Expected behavior observed when a non-pd.Timestamp column is present in the dataframe

import pandas as pd

print(pd.DataFrame([dict(
    b=pd.Timestamp('2023-01-01 12:00:00'),
    c=pd.Timestamp('2023-01-01 12:00:00'),
)]).values.tolist())

# [[Timestamp('2023-01-01 12:00:00'), Timestamp('2023-01-01 12:00:00')]]

Installed Versions

INSTALLED VERSIONS
------------------
commit           : 965ceca9fd796940050d6fc817707bba1c4f9bff
python           : 3.9.6.final.0
python-bits      : 64
OS               : Darwin
OS-release       : 21.6.0
Version          : Darwin Kernel Version 21.6.0: Thu Mar  9 20:08:59 PST 2023; root:xnu-8020.240.18.700.8~1/RELEASE_X86_64
machine          : x86_64
processor        : i386
byteorder        : little
LC_ALL           : None
LANG             : en_US.UTF-8
LOCALE           : en_US.UTF-8

pandas           : 2.0.2
numpy            : 1.24.3
pytz             : 2023.3
dateutil         : 2.8.2
setuptools       : 58.0.4
pip              : 23.1.2
Cython           : None
pytest           : None
hypothesis       : None
sphinx           : None
blosc            : None
feather          : None
xlsxwriter       : None
lxml.etree       : None
html5lib         : None
pymysql          : None
psycopg2         : None
jinja2           : 3.1.2
IPython          : None
pandas_datareader: None
bs4              : None
bottleneck       : None
brotli           : None
fastparquet      : None
fsspec           : None
gcsfs            : None
matplotlib       : None
numba            : None
numexpr          : None
odfpy            : None
openpyxl         : None
pandas_gbq       : None
pyarrow          : 10.0.1
pyreadstat       : None
pyxlsb           : None
s3fs             : None
scipy            : None
snappy           : None
sqlalchemy       : 1.4.47
tables           : None
tabulate         : None
xarray           : None
xlrd             : None
zstandard        : None
tzdata           : 2023.3
qtpy             : None
pyqt5            : None

[topper-123]

Your expected behavior matches actual behavior, can you clarify what think is wrong. currently. Also. Pandas has only control of what's happening until df.values. Thetolist method is a numpy method and is outside of pandas' responisibility,

Looking briefly, everything looks to be working as expected here on pandas's side. The second example returns a numpy array of dtype dtype='datetime64[ns]', which looks right, while the first returns an array with object dtype, which also look alright.

[aminnj]

hi @topper-123. I made a copy-paste mistake! The expected behavior should be

import pandas as pd

print(pd.DataFrame([dict(
    b=pd.Timestamp('2023-01-01 12:00:00'),
    c=pd.Timestamp('2023-01-01 12:00:00'),
)]).values.tolist())

# [[Timestamp('2023-01-01 12:00:00'), Timestamp('2023-01-01 12:00:00')]]

(also edited above)

[aminnj]

Given the edit, can you clarify if this is still on the numpy side?

[topper-123]

Your dataframe columns have the same dtype:

>>> df = pd.DataFrame([dict(
...        b=pd.Timestamp('2023-01-01 12:00:00'),
...        c=pd.Timestamp('2023-01-01 12:00:00'),
... )])
>>> df.dtypes
b    datetime64[ns]
c    datetime64[ns]
dtype: object

So naturally df.values will give an numpy.array of that type, seems logical to me. In your first example the dataframe dtypes were not the same, so df.values gives an array of object dtype.

So IMO this is not a bug and works as expected. I'm closing, but if you post a follow-up, I'll respond and reopen if necessary.