translation workbench is a dumpster fire

i just spent an hour trying to get an import file working, because translation workbench…

the documentation ON THE IMPORT PAGE says

Include the language attribute at the top of the file. Your format should look like Language = “en_US” and include a valid language reference.

proof:
proof.png

Nice, if it was true, but its complete fabrication. It makes not a single iota of difference if it is there or not or filled with garbage.

What is required, is the following

Language code: [valid language code]
Type: [source | bilingual | data]

I spent a freaking hour trying to do shit with Language = xyz, making sure my file is utf8 etc etc etc. What garbage.

So this worked fine

Language = "dumpsterfire"
Language code: es
Type: Source

# KEY   LABEL

CustomLabel.FD_GAR  I am a Sucker

magic.png

 
2
Kudos
 
2
Kudos

Now read this

Preventing duplicate junction object records

Native dupe management only allows you to reference a single lookup field. So it cannot prevent duplicate junction object records. By using a (new in spring 20) before save flow to concatenate the two IDs into a single text field, you... Continue →