How can we know the correct setting without any data? The following is a typical row from my one of my German bank accounts copied from a plain text editor.
10.01.2022;10.01.2022;"Dauerauftrag Ăberweisung";"Stadtwerke xxxxxxx";"VKto 2135004 Stadtwerke xxxxxxxx Info: VKto 2135004";EUR;;579;"759,78"
The fields are separated by semicolon.
First 2 columns are German dates DD.MM.YYYY
Last 2 columns are comma decimals.
Some strings are double-quoted. The EUR is unquoted. Surprisingly, the last comma decimal is quoted too. This is a rather silly flavour of csv.
That bank encodes text with Windows 1252. With wrong encoding, the umlauts (âĂberweisungâ) would be wrong.
I import this row set with import locale âGerman(Germany)â which takes care of all the number formats (including German month names in dates if there were any).
I check âDetect special numbersâ which is the right choice in 99% of all use cases. Otherwise it would import any dates as strings. Unfortunately, this is off when you see this dialog for the first time.
I untick âQuoted fields as textâ because there is one field with a quoted number to be imported as a number.
I specify the explicit field type in the preview section only if some csv requires an exception to the above rules. I want to import phone numbers, zip codes or priduct IDs as text. Sometimes you have US dates with comma decimals or non-Dollar currencies.
The settings are saved, so I only have to confirm the dialog when the next csv is a similar one.