| converting tbx 2 csv | Nov 6, 2011 |
I, too, wanted to convert the microsoft termbase collection to csv (like the old versions) for proper import to trados multiterm (2009). Multiterm's facility does not import all the fields, i.e. definition, part of speech, only the actual terms. This is enough for my purposes, but I wanted the other fields for completeness. Because the document definition is missing (the .xsl file), they cannot be seen and edited by other programs. It opens OK in excel, but because of the missing dtd does not li... See more I, too, wanted to convert the microsoft termbase collection to csv (like the old versions) for proper import to trados multiterm (2009). Multiterm's facility does not import all the fields, i.e. definition, part of speech, only the actual terms. This is enough for my purposes, but I wanted the other fields for completeness. Because the document definition is missing (the .xsl file), they cannot be seen and edited by other programs. It opens OK in excel, but because of the missing dtd does not line up all the fields (only source and target), the definition and part of speech fields are offset down by 2 lines. A laborious process to correct manually (the only solution I've found so far).
What works:
1) open with excel (2003 and later, or any spreadsheet that can deal with xml) and extract the two columns into a new sheet, edit to your heart's content and save) and then import that file into multiterm 7 and later or indeed any other terminology manager. Multiterm 3 and 5 require a slightly different process. You can add other languages from the other downloaded files, if you synchronise the other languages onto English, and, if you want, remove English altogether once finished.
2) use xbench to get the csv, again only the source and target fields.
3) use excelling multiterm, again only the source and target.
I have not tried using microsoft word to open the tbx file and then convert to tab delimited table, because I only thought of it right now.
I also seem to remember that old versions of wordperfect (6 and 8) had the facility to deal with sgml files and consequently xml files, which might be of interest to users of IMB tanslation manager or OpenTM and might make editing the tbx and saving in a glossary importable format a lot easier, i.e. without the use of mail merge and postscript to create the sgml file.
Going back to the original post, it would seem that a tbx file does not include the martif definition, or a martif definition must be created, if there are more than 2 fields 9clomns). ▲ Collapse
| | |
Login to reply/comment 
