← Back to context

Comment by herbst

2 days ago

I don't know any of these problems. I use a modern operating system and office suite that supports CSV not a specific subset and syntax of it.

The syntax that MS Office uses to read/write a CSV is defined by the Regional Settings of your PC.

Open control-panel for regional settings, select "Advanced settings" button on the bottom control.exe intl.cpl

If you don't know any of these problems, then all the people and systems you work with have a "." as decimal and "," as separator, and you are spared from the hell of MS Office being unable to overrule these OS-settings when treating a CSV

  • Honestly as this always was an obvious issue I usually just used ; and never got a complain. Obviously both . And , are used way to often not only for numbers. I am surprised this is problem enough (in 2025) that people emotionally discuss it.

    • > Honestly as this always was an obvious issue I usually just used ; and never got a complain.

      Thing is, it is not about what you used, you are not able to control this from happening when your CSV should work for people in other countries. Whatever configuration you used which never got a complain, if your recipients also used Excel to work with those documents, they probably have the same regional setting on Windows for list/thousands/decimal separator.

      If you use ";" as separator, i.e. Excel in UK, US, Japan, China, Korea will not be able to correctly open your CSV.

      But even better: If you created this CSV on a France or Sweden regional setting, the thousands separator will be a whitespace ("1 000" instead "1,000" or "1.000"), so Excel in e.g. Italy will not detect those properly.

      > I am surprised this is problem enough (in 2025) that people emotionally discuss it.

      It is a (intentional) weakness of MS Office for those who work in an international environment, because Excel links itself to .csv files to hinder the experience, as it is neither able to properly detect them nor guide their users through a process to properly handle them.

1.01 in US === 1,01 in EU

   1.01, "hi", CSV has problems, "1.01"
   1,01, "hi", Yes it really does, "1,01"

See the problem now?

Your operating system cannot solve this problem.

  • CSV already solved this problem with quotes. Maybe not the most convenient solution for some users but that's no excuse for the Excel behavior of making up a different format depending on the locale.

    • Excel really doesn't care what users think. I mean, in biology, we've already had to change the names of genes to accommodate Excel's auto-date conversion routines. So, why would it care to have globally consistent CSV formats?

  • Is this 2025? Why would any software safe it invalid like that to begin with?