Translation memory (TM) – most common issues & troubleshooting


Below are the most common problems clients report in tickets concerning translation memory (TM).

Before reading the article below, we recommend that you read the step-by-step guideline to importing your TM: How to import a translation memory (TM) into XTM.

Before reading the article below, we recommend that you read the step-by-step guideline to importing your TM: How to import a translation memory (TM) into XTM.

Issue description & Solution

TM import cannot be performed because of the size constraint.

This is one of the most common issues a user who is importing a TM can encounter in XTM Cloud. The largest file size you can upload to any server is 150MB. If your file exceeds this limit, the import will fail and an error will be generated.

Solution: There are a couple of solutions you can implement to resolve this:

  • Divide your file into smaller chunks,

  • Zip your files and upload the zip archive,

  • Delete unnecessary parts of the file, especially duplicates, to make it 'slimmer'.

TM import cannot be performed because of unescaped entities

This is the second most common problem encountered by our clients.

Solution: When uploading a TM file, make sure that entities that are not supported by XML standards are properly escaped.

For example, characters like & (ampersand) must be entered as &.

If characters of this type are not correctly escaped, the file will fail to import and the process will fail, either partially or completely. You might find online syntax checkers helpful when you check if your file complies with character standards. See an example of this kind of site: Validate XML files.

TM import cannot be performed because a file format is unsupported

A third major contributor to issues with TM is an unsupported file format or faulty structure.

Solution: The currently supported TM file formats are as follows: TMX, XLIFF, XLS.

See attached a zipped file containing each sample file for its specific format:

"&" or other special characters being displayed in the TM source

This can happen after uploading a TM alignment file in which source segments display &amp marks, whereas the required format would be &.

Solution: Follow the steps below:

  1. Export affected TM (click here for help);

  2. Delete affected TM;

  3. Perform the "Search and Replace" option in the exported TM document and replace &amp with &;

  4. Reimport the corrected TM (click here for help).

Can the deleted TM be restored?

Solution: If you have deleted a TM by mistake, the XTM International Support team can only restore it if the following option was enabled on your account prior to this incident (administrative privileges required!): Configuration → Settings → Translation → TM → Matches - general → Save TM in segment history.

If this is the case, create a support ticket for the XTM International Support team and provide the details.

If the deleted TM is only for one or two projects, keep in mind that you can always retrieve it directly from the XTM Project.

Go to Project editor → Files → TMX, and generate the file in question. You can then seamlessly use it for reimporting the TM from this project.

The file is generated live from whatever is found in XTM Workbench at a particular time.

Batch change of TM status

If you find that your TM has Approved or Not approved status in the TM database, contrary to your expectations, the XTM International Support team can perform a batch database change for that TM.

Solution: Create a support ticket for the XTM International Support team and provide the following information:

  • Customer that the TM is connected to,

  • Language combination,

  • Specify if the change only needs to be performed for the TM related to a specific project or projects (provide their names or IDs. Click here for help.) or all of the projects with this particular customer and language combination.

What is SYSTEM:USER: AUTO-STEP that appeared in the place of user details?

Solution: "SYSTEM:USER: AUTO-STEP"; if this occurs anywhere in the TM-related part of XTM Cloud, this means that a particular match was created by an automated step or process. Typically, it is the Approve TM action or a bulk action such as a Rollback that is marked this way. In the cases above, this is the expected behavior.

Records cannot be found in the TM tab , even if they do exist in XTM Cloud and are displayed as a match in the editor

Sometimes, you might not be able to search for records for single or multiple languages in the TM Manager tab.

Solution: In this case, we recommend that you contact the XTM International Support team as soon as possible to find the cause of this problem. This is usually related to damaged translation memory indexes, which might have been damaged during a sudden or unexpected downtime or if the XTM Cloud services restart. In such a case, the TM needs to be reindexed.

Before creating a ticket, however, make sure you have a good understanding of what TM reindexing is and what details to include on the ticket form. You will find this information in the relevant article: Reindexing of translation memory.