Importing translation memories

The following instructions demonstrate how you can import existing translation memories (TMs) in the TMX, TTX and GLO formats.

The crossTank Import Wizard helps you to import your existing TM databases and assign them to the respective languages and subjects in crossTank.


Before importing translation memories to crossTank, we recommend optimizing the number and sentence detection settings for the languages to be imported. For example, if the definitions of the number formats in Across correspond to the number formats in the translation memory to be imported, the numbers will be identified correctly during the import, resulting in better and more consistent data.

The following example demonstrates the import of translation memories from a TMX file. The import of other formats TTX as well as GLO takes place in the same or at least similar way. So you can perform the import for these file formats in an analogous manner.

  1. Proceed as follows:
  2. Open the crossTank Manager.

    If you open the crossTank Manager from within crossDesk, the settings of the project that the task belongs to will be used.

    If you open the crossTank Manager outside crossDesk, the settings of the default project settings template will be used.

  3. Start the crossTank Import Wizard via the corresponding section on the crossTank Manager start page, via the icn_wiz_cTank-import-wizard icon in the toolbar, or via Tools > Import in the menu bar.
  4. Once the wizard has started, click Next.
  5. Select the format you want to import and click Next.
  6. Select the file you want to import.
  7. Select the mode with which you want to import the file.
    • Import with user interaction
    • Import without user interaction
  8. If you created an import template (*.import) during a previous import of a TMX file, you can select it now.

    Use an import template only for another import of the same data as during the creation of the import template or for an import of other data with the same attributes and attribute values.

  9. Finally, you can determine that Across is to perform auto-adjustments during the import.
  10. Click Next.

    The selected file is parsed, i.e. the file is read out and analyzed.

  11. Activate the respective checkboxes to select the languages whose translation units you want to change.
  12. Click Next.

    Now the language mapping takes place. The languages and sublanguages of the translation units from the import file are assigned to the languages and sublanguages in Across.

  13. Click the corresponding language in the Across language column to map the desired language. Select a language from the drop-down list. Proceed in a similar way to map sublanguages. Continue until all languages and sublanguages you want to import are mapped. The languages from the import file that were not mapped to any language will not be imported.
  14. Click Next.

    If the file you want to import contains translation units consisting of more than one sentence in the source segment and/or target segment, the following dialog is displayed.

  15. You can determine whether the translation units containing more than one sentence are to be imported as a single segment (without segmentation) or split in sentences as multiple segments (with sentence segmentation).

    The selection depends on the segmentation method selected under Tools > System Settings > General > Segmentation.

  16. After selecting the needed settings, click Next.

    Now the mapping takes place.


    The date and time of creation and modification of TM entries will be adopted when the TMX file is imported. If the TMX file does not contain this kind of information, the import time will be set for the TM entries.

  17. Continue until all attributes of the import file are mapped to the respective Across attributes, then click Next.

    Please note that information in attributes of the import file that are not mapped to any Across attribute will be lost during the import process.

    Next, the mappings that you have made will be specified. The attribute values of the translation memory file are mapped to the corresponding attribute values available in Across.

  18. To do this, proceed as you did in the previous step. Select the column of an attribute value and then use the drop-down list in the column on the right to map the appropriate Across attribute value.
  19. Continue until all TMX attribute values are mapped to the respective Across attribute values, then click Next.

    Information contained in TMX attribute values that are not mapped to any Across attributes will be lost during the import process.

  20. During the import of a TMX file, click Save import template to create a template with the settings you selected for the import (and a file extension *.import).
  21. The storing settings configured for the import are displayed under Storing settings. These settings concern the storing of attributes and of multiple translations. Click Storing settings in order to change the settings for the import.
  22. Click Start import.

    The import of the translation units begins.

    If you selected Show manual correction dialog, dialog windows for manual corrections will appear during the import.

  23. While importing the translation units, Across checks the numbers they contain on the basis of predefined rules. If the source segment and the target segment contain different amounts of numbers, a dialog window indicating this will be displayed. Click Store translation to import the translation unit to crossTank or click Edit translation to edit the translation unit before importing it.
  24. Click Finish to complete the import.

Upon completion of the import, click Show report to display an overview of the most important data of the import, such as a summary of the performed mapping operations and the number of imported translations. Click the icn_report-drucken icon to print the report. Use the icn_report-zwischenablage icon to copy the report to the clipboard. Click the icn_report-speichern icon to save the report as an HTML document.

The translation units contained in the TMX file are now fully imported in crossTank and can be used in Across; the imported translation units have the status Smartly inserted (icn_match_smart-eingefuegt).

If you wish, you can now edit or check the new translation units in the crossTank Manager and change their status from Smartly inserted (icn_match_smart-eingefuegt) to Released (icn_match_freigegeben).

Manual Corrections

Translation units with the same number of source and target sentences:

If the mapping of source and target sentences is correct in the window below, click Save matches to transfer the translation units to crossTank and continue the import.

If the mapping has not been done correctly, mark the respective sentence pair(s) and click Remove match. Click Join all to save all sentences of the paragraph pair in a single segment. You can also align the sentences correctly by selecting the corresponding sentences in the source and target segments in the top left and right panes and click Add match. Continue until all sentences are matched correctly and click Save matches in order to transfer the translation units to crossTank and continue with the import.

If you would not like to make any corrections manually and discard the translation units, click Cancel.

Translation units with a different number of source and target sentences:

Click Join all to save all sentences of the paragraph pair in a single segment. To correct the translation unit(s) manually, select the sentence(s) in the source and target segments that belong to one translation unit and click Add match in order to join the sentences to one translation unit. Continue in the same way until all segments have been mapped.

Click Save matches to transfer the translation unit(s) to crossTank and continue with the import.

To skip the manual correction and discard the translation segments, click Cancel.