It seems the delimiter autodetect is very easily tricked when multiple of the autodetect characters are present, even if just once. With the default settings (both ; and , set up), the following (fake but realistic) data will split at the comma (,) instead of at the semicolon (;), even though the semicolon is all over whereas each line only has one comma:
April 15, 2023;4028722987;25.00;4.00;16.00;Ray;pending;
April 14, 2023;4441335566;25.00;4.00;16.00;Scott;pending;
April 14, 2023;1012167625;25.00;4.00;16.00;Paige;pending;
April 13, 2023;5595193937;25.00;4.00;16.00;Antone;pending;
April 12, 2023;3329268556;25.00;4.00;16.00;Theron;pending;
April 11, 2023;7187180132;25.00;4.00;16.00;Brandy;pending;
April 10, 2023;1327553107;25.00;4.00;16.00;Lupe;pending;
April 9, 2023;6783434534;25.00;4.00;16.00;Lonny;
April 9, 2023;6387478020;25.00;4.00;16.00;Rosa;
April 8, 2023;7795581629;25.00;4.00;16.00;Cecelia;
I would think it should be looking at either the last character or the most appearing (non-escaped) character, prior to assuming it's supposed to be the first appearing delimiter instead.
Apparently it does work if the column lengths are the same. Somehow I had a database being exported with an additional semicolon in the end (or rather trailing semicolons being trimmed from the other rows).
It seems the delimiter autodetect is very easily tricked when multiple of the autodetect characters are present, even if just once. With the default settings (both ; and , set up), the following (fake but realistic) data will split at the comma (,) instead of at the semicolon (;), even though the semicolon is all over whereas each line only has one comma:
I would think it should be looking at either the last character or the most appearing (non-escaped) character, prior to assuming it's supposed to be the first appearing delimiter instead.