ENH: Different behavior of pandas when saving and restoring from a CSV file

[pevogam]

Is your feature request related to a problem?

It seems that df.to_csv(filename) will automatically save the index (because of the choice of default) while pd.read_csv(filename) will then automatically add its own index (again because of the choice of default). A user relying on simple default behaviors doing multiple reading and writing to an CSV file will end up accumulating indices:

      Unnamed: 0  Unnamed: 0.1  Unnamed: 0.1.1  Unnamed: 0.1.1.1         p       q    r
0              0             0               0                 0  54.78  0.0005  1.0
1              1             1               1                 1  54.78  0.0005  1.0
2              2             2               2                 2  54.78  0.0005  1.0
3              3             3               3                 3  54.78  0.0005  1.0
4              4             4               4                 4  54.78  0.0005  1.0
...          ...           ...             ...               ...       ...     ...  ...
2360        2360          2360            2360              2360  54.78  0.0005  0.0

and thus can get tripped by this difference in choice of defaults.

Describe the solution you'd like

I would recommend settling on the same default behavior when both storing and retrieving from CSV, possibly simply not storing the index by default when using to_csv to obtain a symmetric result.

API breaking implications

I assume this might break the current API unless we provide a warning for the ongoing change at least for a couple of versions. After all, API is also not meant to be frozen and never improved just because it is being used.

Describe alternatives you've considered

None, this is just a suggestion.

Additional context

Possibly added above.

[phofl]

Duplicate of #24468

[jreback]

see #4595 and there are some others

long discussed and rejected

[pevogam]

Yes, I will monitor the issue this one turned to be a duplicate of. Thanks for the hints!