add concat function dtype test

[liang3zy22]

• closes Incorrect dtype for concat with empty float and datetime columns #32934
• All code checks passed.

The old PR closed, so created new one to close #32934.

[liang3zy22]

1 check failed with current code,Ubuntu/Data Manager. It means that the code
pd.concat([df_time, df_float.iloc[:0]])
has data type "object".
If I updated the codes to the following code:
tm.assert_equal( pd.concat([df_time, df_float.iloc[:0]]), pd.DataFrame({"A": pd.array(["2000"], dtype="datetime64[ns]")}).astype({"A":"object"}), )
The Ubuntu/Data manager will pass, but other 21 checks will fail since the code
pd.concat([df_time, df_float.iloc[:0]])
may have data type "datetime64" in these 21 checks(I only checked some).

[mroeschke]

For each of these:

1. Could you use tm.assert_frame_equal
2. Create variables expected = ... , result = ...

[mroeschke]

Looks like the using_array_manager fixture needs to be used to set the dtype for the last example

[mroeschke]

Could you move this test to pandas/tests/reshape/concat/test_datetimes.py? After that this looks good

[liang3zy22]

It seems that another numpy error in "Ubuntu / Numpy Dev(pull_request)".

>       warnings.warn("`product` is deprecated as of NumPy 1.25.0, and will be "
[134](https://github.com/pandas-dev/pandas/actions/runs/4379948504/jobs/7666450861#step:8:136)
                      "removed in NumPy 2.0. Please use `prod` instead.",
[135](https://github.com/pandas-dev/pandas/actions/runs/4379948504/jobs/7666450861#step:8:137)
                      DeprecationWarning, stacklevel=2)
[136](https://github.com/pandas-dev/pandas/actions/runs/4379948504/jobs/7666450861#step:8:138)
E       DeprecationWarning: `product` is deprecated as of NumPy 1.25.0, and will be removed in NumPy 2.0. Please use `prod` instead.

[mroeschke]

Thanks @liang3zy22 (that failure is unrelated)