• [ x] I have checked that this issue has not already been reported.

• [ x] I have confirmed this bug exists on the latest version of pandas.

• [ x] (optional) I have confirmed this bug exists on the master branch of pandas.

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

for k in range(9,17):
    dstr = '{}-07-2021 15:57:40'.format(k)  # assumed format ''%d-%m-%Y %H:%M:%S' eg. in ch/DE
    print(' dstr: {}  pd.to_datetime(): {}'.format(dstr,pd.to_datetime(dstr)))

Output:
dstr: 9-07-2021 15:57:40  pd.to_datetime(): 2021-**09**-07 15:57:40
 dstr: 10-07-2021 15:57:40  pd.to_datetime(): 2021-**10**-07 15:57:40
 dstr: 11-07-2021 15:57:40  pd.to_datetime(): 2021-**11**-07 15:57:40
 dstr: 12-07-2021 15:57:40  pd.to_datetime(): 2021-**12**-07 15:57:40
 dstr: 13-07-2021 15:57:40  pd.to_datetime(): 2021-07-13 15:57:40
 dstr: 14-07-2021 15:57:40  pd.to_datetime(): 2021-07-14 15:57:40
 dstr: 15-07-2021 15:57:40  pd.to_datetime(): 2021-07-15 15:57:40
 dstr: 16-07-2021 15:57:40  pd.to_datetime(): 2021-07-16 15:57:40

Problem description

It would be ok if pd.to_datetime() assumed a default for parameter format, for example either US-format, or ISO-format '%Y/%m/%d %H:%M:%S'.

Anyhow as the code sample show, to_datetime() changes it's assumptions based on each input string. This is not acceptable and can be very harmful. There is no good in such dynamic parsing strategy.

I guess pd.to_datetime() assume a dateformat of '%m-%d-%Y %H:%M:%S', once this assumption is violated, the decoding strategy changes to

'%d-%m-%Y %H:%M:%S' instead of throwing an exception. Such instability is no good. pd.to_datetime() should work according a given date format. First violation according that date format should throw an exception

Expected Output

Any time pd.to_datetime() should work under control of a format string, a datetime string format. This could either be a default or a parameter passed format string.

If no format string is passed as a parameter to_datetime() uses a flexible strategy. As Pandas main scope is to work with DataFrame and Series object, timeseries are a prominent application. Such flexible date format interpretation can here be very destructive. Such behaviour of to_datetime() should not need to be expected and is NOT failsafe.

Output of pd.show_versions()

INSTALLED VERSIONS

commit : 7c48ff4
python : 3.8.8.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19041
machine : AMD64
processor : Intel64 Family 6 Model 45 Stepping 7, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : German_Switzerland.1252

pandas : 1.2.5
numpy : 1.20.2