Remove tag codes from Numeric Mismatch


When running QA, I noticed that certain tag numbers appear as Numeric Mismatch.
Is there perhaps a Regular Expression or something we can add to the Checklist to modify it?

For example:
The day before you are discharged is your last inpatient day<g id="292">.</g>
El día antes de recibir el alta es su último día como paciente internado.

Also, in the Tag Mismatch instances, for example if there is an italic tag, it appears differently in the source and target in Xbench causing this to display (1336 on top and Italic on the bottom):

For information about your costs in these stages, look at your <g id="1336">Summary of Benefits</g>
Para obtener información acerca de los costos en estas etapas, consulte su <g id="Italic">Resumen de beneficios</g>

Thank you very much,


This happens in Trados Studio when the source tags are not applied to the target text even though text format is the same as in the original file.

In order to the display formatting tags in Studio 2017, you should go the View tab and enable Toggle formatting tag display and Full Tag Text.


Thank you, but is there a way to prevent Xbench from seeing these as Numeric Mismatches?


Go to the QA tab and enable the Ignore Tags checkbox at the Options section.

In this way, Xbench will omit numbers inside tags from Numeric Mismatch.


@leah_spielholz, in any case, please keep in mind that apparently innocuous mismatches in <g id='xxx'> tags may have unwanted side effects, such as not being able to export the target Word file in SDL Trados Studio.


Thanks omartin. This worked.
Anyway, this is one of the worst improvements they have done to Xbench. I’m still thinking of finding a way to downgrading. This function is obviously useless as you can get the same results from tag mismatch check, plus it makes it hard to find any real numeric mismatches when you have to go through all these irrelevant results.

If somone from Xbench team is reading this, please give it a thought.


We will consider not checking numeric mismatches within a tag, because as you say, the tag mismatch should already report it.

In the meantime, does it help if you go first through the tag issues and once all of them are fixed then just press F5 to refresh updated files, click Check Ongoing Translation again and proceed with with numeric mismatches?

Everything happens on the QA tab so, in normal circunstances, we are talking about a matter of just a few seconds to press F5 and then Check Ongoing Translation for the regression.