CSV imports don't separate columns by commas

I was updating the translation files for a project I’m working on just now, and I noticed that all the translations were deleted due to a wrongly imported CSV file. After examining Crowdin’s interpretation of the new file’s table, it seems that it didn’t recognize and separate columns by commas.

For example, it would consider this as one column, and places it on the first row:

Key,Id,Chinese (Simplified)(zh-Hans),Chinese (Traditional)(zh-Hant),English(en),Japanese (Japan)(ja-JP),Korean(ko),Vietnamese(vi)

I then used an online CSV viewer and it was able to identify and separate columns just fine.

This problem only occurred to me today since I was able to update that specific file with the commas just fine a while back. Are there any solutions / workarounds?


Hello, 2 assumptions:

  1. Wrong language codes, try checking this link and compare yours Language Codes | Crowdin Developer Portal
  2. If you use some automated tool, probably this configuration may help Configuration File | Crowdin Developer Portal
    2.1) With 1st UI import you should do that manually CSV / XLSX Files Configuration | Crowdin Documentation

Thanks for helping!

The language codes on that project are set up manually via Crowdin’s import configuration, and it’s worked before, so I think that’s good.

By the way, this is what comes up when I try to update a table and look at its schema:

Hi, tried to recreate your case, created a CSV in Libreoffice tables with your set up

Key,Id,Chinese (Simplified)(zh-Hans),Chinese (Traditional)(zh-Hant),English(en),Japanese (Japan)(ja-JP),Korean(ko),Vietnamese(vi)

It doesn’t recognize spaces. Try to remove them, so it would be without spaces after/before the commas, and in that case all would be updated

1 Like

Jumping here, @2394425147 I’ve gone ahead and activated revision history for your account from my side, you can track changes within that file by downloading previous versions. More than sure that you will find commas as a difference.

1 Like

Thanks for the advise, I think it’s working again!

Turns out, in the updated table, I had extra columns in the last row, and I think the difference between the layouts messed it up. After cleaning the cells, Crowdin was able to identify the columns and rows as usual!

Another thing I looked over was that I use Unity’s localization package (v1.0.5) to export the tables, and it somehow introduces corruption when trying to overwrite existing files, so that might be it.

Link here Discover the right solution to enhance Crowdin experience redirects to higher versions like 1.2 and 1.4 About Localization | Localization | 1.2.1

Probably this is the point as well