Importing bilingual tables to a TM

Many ways to go, some more sensible than others

I’ve seen questions like this so many times, and so many times the answers I see are unnecessarily complicated for the situation. And too often, in responding to such a question, the ones providing suggestions are distracted by irrelevancies and possibly false assumptions of the questioner.

The essence of the question is highlighted in the text. (I copied it from a social media platform into Microsoft Word just so I could do the discontinuous highlighting.) And the answer would apply to pretty much any translation environment tool I’ve seen. The rest is of no importance, just opportunities for some to end up in some strange solution space in a dimension of little practical use.

Someone has a two-column table in a word processing file. The table may actually have more than two columns, but what is important is that one of these columns contains a Language A, and the other column contains the translation of the first column into Language B in the same row. Something like this:

Data from the EUR-LEX website

or this:

A bilingual table export in a word processing file… an option in many CAT tools today

Now how would you get data like that into a translation memory?

Before you start going wild with notions of saving the source and target columns in separate files, or even transferring the data to Excel, consider for a moment that most translation memory systems can import delimited text files. Does that information change the approach you would take?

It should. Microsoft Word (and other word processing applications) can convert tables to text, typically tab-delimited text. So do that:

And then import the tab-delimited text to your translation memory:

And here’s the result in the memoQ TM editor, though of course this approach would work with nearly any other translation workspace tool:

Of course there are situations where certain formatting in the table may require additional preprocessing (deleting columns, dealing with multiple paragraphs in cells, etc.) but for the case of a table export for review from an application like Trados Studio, Fluency or others with sentence-level segmentation comprising the table cells, it will work without trouble. If you are a memoQ user receiving a memoQ RTF bilingual table export, you can simply import this to a project, attach the translation memory to which you want to write the content as the Working memory, and then confirm all the segments again to write them to that TM.

Users of translation environments other than memoQ can use this approach of saving to delimited text (after deleting the header and the ID column) to get memoQ RTF bilinguals (or any other tool’s table exports) into their tool’s TMs.