Multiterm Converter 2021 Parsing Error

Question

I am trying to convert a .tbx file to create a termbase. 
 The .tbx file is from the Irish terminological database, available here https://www.tearma.ie/ioslodail/ 
 I am getting this error: 
 I can see other parsing errors on the forum, but nothing that is a help re Multiterm Converter 2021. Can anybody advise? 
 TIA, Dar&aacute;n

Oana Nagy · Accepted Answer

Hi @ Dar&aacute;n &Oacute; Dochartaigh , 
 You can do the same, converting a .tbx file into a termbase .sdltb with the Glossary Converter, in fact it&acute;s easier and more intuitive to handle, please check it: 
 https://appstore.sdl.com/language/app/glossary-converter/195/ 
 And you can check here how to use it: https://www.youtube.com/watch?v=vYuSiN1bVfQ&t=1s 
 Let me know if it works alright this way, please! 
 Regards,

Former Member · Answer

I downloaded this tbx file, changed the extension to xml and then used the SDL MultiTerm 2021 Convert (selecting OLIF XML format). That error refers to an ampersand (that is, the & sign) which is mixed up with the actual content of the terms. The ampersand cannot be escaped (ie replaced) with \& and instead Unicode &#038; should be used. However there are tons of error of the same type plus other ampersands mixed up with backslashes. This termbase has 184420 entries (ie "concepts") and several descriptive fields (inside 4165212 lines). Doing a manual clean up is very slow and I guess a global search and replace with regular expressions will do the job. I haven't tried with the Glossary Converter, but it is most likely the same errors will occur. I used Notepad++ to explore this huge XML file (in addition, files with the extension .tbx are essentially xml files, so changing the extension may not be necessary).

Former Member · Answer

I could not make it work using the Glossary Converter, which in addition highlighted other problems. The XML is validated OK and it's well formed. In my view, handling correctly the ampersand, which is interspersed in the actual content, will be the solution.

Paul · Answer

Darán Ó Dochartaigh 
 I downloaded the TBX from the website. As noted the file does contain invalid XML. So this is what I did: 
 
 ran a couple of search & replace operations on the file to remove all the invalid XML and replaced with &apos; which is the correct way to represent the reserved character for an & symbol in XML. As far as I can see there are no other reserved characters being used incorrectly 
 I then converted the TBX (it's not OLIF) using the Glossary Converter rather than MultiTerm Convert because it's easier ;-) There were a few entries in Latin (LA), Maori (MA), and old Irish (SGA), none of which are supported in Trados so I ignored these 
 the entire TBX contains a large number of languages which makes it much too large for MultiTerm to convert and handle. So as a proof it works I ignored all languages apart from English and Irish and this works... just takes so long to reorganise I lost patience and didn't wait till the end... but did get a good number in just to prove the concept: 
 
 Reality is that aside from the XML being invalid it's simply too large! 
 https://multifarious.filkin.com/2019/06/29/how-do-you-eat-an-elephant/ 
 Also noting... "concepts" and "entries" are not the same thing ;-)

Paul · Answer

Darán Ó Dochartaigh 
 It was an interesting test case so I did try to get the entire thing into Trados Live... this worked nicely and it was fast. I had to convert the TBX to an XDT/XML to make this work successfully, but the Glossary Converter makes short work of that once you fix the TBX. 
 I have put the TBX here and also the XDT/XML so you can have a play as you see fit: tearma.zip

Trados Studio > 3. MultiTerm

Multiterm Converter 2021 Parsing Error

Top Replies