Fixing a corrupt TMX file for importing into Across

On occasion you receive Translation Memories in TMX format, in order to add them to your memory for the translation of a project. However, depending on the source of the TMX file, sometimes you will run into an issue where is states that the TMX file is corrupt and cannot be imported. Following you will find a suggested methodology for fixing a TMX file and importing it into Across.

Proceed as follows:

1. Open Across.

2. Switch to CrossTank Manager and select the import function.

3. Select the TMX file.

In this sample you will see that the file has an issue prohibiting the import.

4. Cancel the action.
Open the TMX file with NotePad++ (Freeware editor) and navigate to the line with the issue. You will see that the issue is a formatting character.

However, just deleting the character will not do the trick because it just gives you a new message with the next one. As you can see there are lots of them in here. Therefore continue as follows:

5. Download the Apsic Workbench under http://www.apsic.com/en/products_xbench.html, and install the software.

6. Open Apsic and select Project > New as shown below.

7. In the dialog shown select Add, then TMX file in the dialog for file selection, and click Next.

8. Add the file in Question and click Next as shown below.

9. In the following dialog click OK, to start the processing as shown below.

10. When the file is loaded, switch to Tools > Export Items.

11. Enter the necessary settings and click OK, as shown below.

12. The new TMX file is created in the location specified.

13. Repeat steps 1, 2 and 3. As you can see below, now you have a valid TMX file.

14. Continue as normal to import the TMX file.

Technical comments

The issue described here applies mainly with translation memories in TMX format created with Trados products. This is due to the fact that certain versions of Trados are TMX compliant and compliant does not mean that it is fully supported. So far this procedure has worked with every TMX file that was received with these type of formatting codes included.