How to split .tmx/translation memory by size?

Question

Hi! 
 I would like to split a large .tmx by size into small amounts. How to do that?

Paul · Answer

Olav 
 If it helps... I just used ChatGPT to create a powershell script to do this. Script is here along with the sample TMX I used: 
 https://github.com/paulfilkin/Powershell_scripts/tree/main/split_TMX 
 Also a short video to explain how to use it:

Paul · Answer

Olav 
 Most likely due to content that is in your TMX. I did this in a few minutes and only tested one 33k TU TMX that is probably a pretty good content after coming from the EU. T he errors might be caused by the way the TMX content is handled or by the presence of special characters that are not correctly escaped. Additionally, if the TMX file contains HTML or JavaScript-like content, it might cause parsing issues. 
 If you can't fix this yourself then I'd need to have your TMX to resolve it properly.

Paul · Answer

Olav 
 Thanks, I downloaded it this morning. I don't know if you made a mistake but you sent me a 2.6Gb SDLTM and not a TMX. Anyway, if you tried to use the script on that it would cause a problem! 
 So I exported the SDLTM to a TMX and tested my script. It didn't error at all, but it did fail to extract any segments as it found none. So this prompted me to make a few changes to the script and now it works. I updated it in Github and here's how it works with the TMX created from the SDLTM you provided.

Paul · Answer

Olav 
 A TMX is a flat file so the easiest way is probably to simply use a text editor. Just work out how many you need based on your size requirements, then convert this into how many lines each file when split would contain, and split it up like that. Manual but not difficult. 
 Is that enough of an explanation for you? I don't know how comfortable you might be with xml (a TMX is an xml file) files or working with a text editor.

Caterina Lazzarini · Answer

Hi Olav there are ways to split TMX files into several smaller files. However, you cannot set the exact size of each file, you just set how many files you want the main TMX file to be splitter into, which obviously results in TMX files of smaller size. For example, check this article: 
 https://gateway.sdl.com/apex/communityknowledge?articleName=000001019 
 Another method is to use Heartsome TMX. https://github.com/heartsome/tmxeditor8 Click on: All installer: Microsoft OneDrive, Dropbox (the other links don't work). This is a GPL license so it can be used and distributed for free. Right installer: HSTMXEditor_8_0_1_Win_x86.zip 
 However: For huge TMX files, you might get an error. The error will point you to the solution, which is to increase the memory by changing the Heartsome TMX Editor.ini file and increasing -Xmx512m to -Xmx1024m Once done: Open the tool. In the File menu, you have Split TMX . Choose your TMX file and the number of pieces you want it split into. You can also merge multiple TMX files from the File Menu, by selecting Merge TMX . 
 Note that all these methods are based on non-RWS tools that you can apply at your own risk. We do not provide support for these tools/scripts. 
 I hope this helps you, 
 Caterina

Trados Studio > 1. Trados Studio

How to split .tmx/translation memory by size?

Top Replies