Stepan Konev

I never have a lot of success in trying to correct chars in this way.  If it was me I'd go back to the TMX I can get from Trados 2007, and then make sure the that it is properly encoded.  Only then would I upgrade it or import it into an SDLTM.

Certainly I doubt you'll be able to fix this within the SDLTM, and even exporting it would probably not get you the desired result since the encoding of the new TMX you export is unlikely to be what's needed to support a conversion.

Stepan Konev

This probably also comes to the same conclusion...

https://stackoverflow.com/questions/4532532/check-the-encoding-of-text-in-sqlite#:~:text=You%20cannot%20change%20the%20encoding,And%20then%20create%20your%20database.

I was just curious.

Hi Paul,
Thank you for your reply.
Actually, I wrongly phrased my post... It reads as if conversion spoiled the coding. In fact, the wrong coding was already there in Trados 2007. Sorry for causing misconception.
I could fix it by 66 replacements (33 Russian letters x 2 upper and lower cases). However I found it too time-consuming :) That's why I decided to ask experts. Probably somebody else, as lazy as me, knows how to do it quickly. From other forum I could find a workaround but it involves 3rd-party CAT tool. It can batch replace multiple chars in tmx in one go using a simple UTF-8 txt file with pairs like 
à=а
á=б
...
ß=Я
But I am reluctant to buy a 3rd-party tool for that single task. So far I have a free promo version to fix all my wrongly encoded TMs. And I hope it should suffice for my purposes because this problem does not arise anywhere in newer TMs.
So the problem is solved for me now, but if anybody comes across this thread one day and can suggest a way to batch replace characters, I will still appreciate that.
Cheers,
Stepan

Hi Paul,
Thank you for your reply.
Actually, I wrongly phrased my post... It reads as if conversion spoiled the coding. In fact, the wrong coding was already there in Trados 2007. Sorry for causing misconception.
I could fix it by 66 replacements (33 Russian letters x 2 upper and lower cases). However I found it too time-consuming :) That's why I decided to ask experts. Probably somebody else, as lazy as me, knows how to do it quickly. From other forum I could find a workaround but it involves 3rd-party CAT tool. It can batch replace multiple chars in tmx in one go using a simple UTF-8 txt file with pairs like 
à=а
á=б
...
ß=Я
But I am reluctant to buy a 3rd-party tool for that single task. So far I have a free promo version to fix all my wrongly encoded TMs. And I hope it should suffice for my purposes because this problem does not arise anywhere in newer TMs.
So the problem is solved for me now, but if anybody comes across this thread one day and can suggest a way to batch replace characters, I will still appreciate that.
Cheers,
Stepan

Stepan Konev

ok - so if this simple search and replace will work for you then perhaps use a batch search & replace using the batchedit feature in TM Maintenance?  An example of what's possible is in this KB:

https://gateway.rws.com/csm?id=kb_article_view&sysparm_article=KB0032154

Workaround #2.  You could use this same principle to build a script with your requirements and then run it whenever needed.

Paul said:

You could use this same principle to build a script

I could, but I can't :D I have no idea what pattern to use...

Paul said:

ok - so if this simple search and replace will work for you then perhaps use a batch search & replace using the batchedit feature in TM Maintenance?  An example of what's possible is in this KB

Yes! That's exactly what I wanted. Thank you Sir!