How to export a full Translation Memory to an Excel Workbook 




This export action to an Excel file will take only a few seconds, and even less than a minute for a TM with more than half a million TUs. It all depends on your system settings. You will get clean translation units in plain text, without tags.


A quick note about TM size. It is advisable not to keep TMs larger than 100000 TUs. TMs larger than this are not easily manageable or practical. Large TMs are only useful if you are training Machine Translation (MT) systems or Large Language Models (LLM). You may want to make your TMs as specialized as your own translation service.


The resulting exported file, with the extension .csv, should be first saved as an Excel Workbook (.xlsx). Once you have your complete Translation Memory in Excel format, it is easy to convert it into a TMX file and from there create a Translation Memory compatible with your favorite CAT tool.

      1. In Excel there are two extra "control" columns, B and D. If they are completely empty, you can delete them. Otherwise, for each cell found with content, review the entire Translation Unit. Column A must have only the Translation Units IDs, otherwise review the corresponding TU where there is not a TU_ID. Note: This ID is not related to the IDs in the original Translation Memory in .sdltm format. To explore each column, place the cursor on the top cell, then press the End key, followed by the Down Arrow key.
      2. You can see here the exact number of Translations Units as indicated in the TM module, for this particular example.
      3. The entire TM is exported to a .csv file readable in Excel (we used tab stops instead of commas to separate fields). The resulting file is called 
        TM__Exported by Tb-Scout v3.3__[File name of the TM]__[Date] [Time].csv, which should be saved as an Excel workbook (.xlsx).