We have a FAQ on CSV-import: http://www.webyog.com/faq/28_73_en.html.
In advance you must create a table with the columns 'Commune', 'Codepos', 'Departement' and 'INSEE' of appropirate type.
With these data it would be easiest to 'fill Excel-friendly values' and next edit these two:
'delimited by' is SEMICOLON (Continental European EXCEL CSV-format!)
'enclosed by' is LEFT EMPTY/UNCHECKED
'escaped by' most likely does not really matter here – let it be as \
'lines terminated by' is 'rn' (maybe you know it as – it is just another way of expressing the same thing!) if file is generated on Windows and only 'n' if generated on Linux/Unix.
AT THE BUTTOM you enter “1” so that the header line is not read as data, but skipped.
the only issue that I could foresee here is if you have a ';' (semicolon) in some of your data. That is what escaping is for then. But most likely you don't!