When attempting to finalise one or more target files, error task(s) are produced. Accepting them in your Inbox, and opening them, the error detail shows:
One or more of the segments in the source file(s) are split across two or more lines.
Workaround for Current Projects
If the problematic translation(s) came from the Translation Units (TUs) of one or more of the Translation Memories (TMs), then we recommend exporting the affected language pair(s) to a TMX file, removing any occurrences of
and then recreating that language pair in the TM.
Removing the code from small to medium Translation Memories
Having identified the language pair(s) in the TM(s) that contained the problematic TU, follow the steps below for each one:
WARNING: These instructions should be followed by advanced users only. If you are not confident in performing these steps the please log a Support Case, and we will be happy to assist you.
If your TMX file is very big then Notepad++ may be unable to open it, or your computer may freeze. The solution below may be used in place of steps 2 to 6 above.
awk (UNIX / Cygwin / macOS / Linux)
Where export.tmx is the name of the TMX export, and fixed.tmx is the name of the version with the code stripped:
", "");print}' export.tmx > fixed.tmx
Underlying CauseThis issue occurs because there is a code before the end of the first line of the segment, which is the Unicode hex character for a carriage return (CR). This code is converted into an actual CR and then, before generating the target, Trados Trados Cloud / Enterprise / Live attempts to convert DOS/Windows-type end-of-lines, which use CRs plus line feeds (LFs), into UNIX-type end-of-lines, which are LF only. So, instead of being able to convert CRLF into LF, the affected lines have CRCRLF, and the process fails.
Scenario 1 - Copy / pasted text from other applicationsThis can happen if you copy / paste text from a web browser or a rich text application such as Word, Excel, etc. into one or more target segments in either Trados Studio or Online Editor.
Scenario 2 - Translation came from a Translation MemoryDuring Translation Memory (TM) matching, the affected target segment came from a Translation Unit (TU). In this situation it is sensible to export the TM (or at least the affected language pair) into TMX format, check the entire TMX for other occurrences of
and, if found, to remove them and recreate the TM / affected language pair.
DefectA defect has been logged for this issue, and a fix is expected in a future release of Trados Cloud / Enterprise / Live.
Trados Cloud / Enterprise / Team: How to export cloud translation memory content?
Trados Cloud / Enterprise / Team: How can I delete a single language pair from a Translation Memory?
Trados Cloud / Enterprise / Team: How can I add one of more language pairs to an existing Translation Memory?
Trados Enterprise Online Documentation: Importing TM content