Combine several TMX translation memories into one without duplicates, or split a big TMX by language, year, user or size. Step by step, in your browser.
To merge TMX files, combine their translation units into one file with one header and remove the copies that appear in more than one memory. The quickest way is the free TMX merge tool: drop two or more TMX files, choose whether to remove duplicates, and download one TMX. To go the other way, TMX split breaks one memory into several. Both run in your browser, so nothing is uploaded.
Translators end up with many memories: one per client, one per CAT tool, one per year, an old one from an agency. Merging them gives you one memory to search and one file to move when you change tools. Splitting helps when a client wants only their content back, or when a file is too large for your CAT tool to import comfortably.
Every unit is copied exactly as it was, with its dates, users and properties.
Units without a year, user or property go to a file ending in _unassigned.tmx, so nothing is lost.
Import them into one memory in your CAT tool. Create an empty translation memory in Trados, memoQ or another tool and import each TMX into it, then export one TMX. This is the usual way if you already own the tool. How duplicates and metadata are handled depends on the tool's import settings, so check them before a large import.
Maxprograms TMXEditor. This desktop editor for Windows, macOS and Linux can merge two or more TMX files into one and split a TMX into a chosen number of smaller files, and it has a separate command to remove duplicated units. It needs a subscription after a free trial.
By hand in a text editor. A TMX is XML, so you can copy the body of one file into another. It works for two small files with the same languages and encoding, but you get no duplicate removal and it is easy to break the XML.
For the formats involved, see XLIFF, TMX and TBX explained. If you are merging because you are leaving a tool, the memoQ alternatives guide covers moving memories between tools.
Duplicates that aren't removed. Merge removes units whose text and tags are identical in every language. Units that differ only in a double space, a tag or the punctuation are different, and units with the same source but a different translation are variants, not duplicates. Run the merged file through the TM cleaner to fix spaces and choose between conflicting translations.
Language codes that differ. One memory may say es-ES and another es. Codes are compared ignoring capitalization, but es and es-ES are different languages, so a one-pair merge for es-ES leaves out units marked only es. Check the codes before merging.
Different source languages. If one memory is English to Spanish and another Spanish to English, the tool lists the file with a different source language. The merged file uses the first file's header, so decide which direction you want before merging.
The wrong header. The first file's header is kept, including its source language and the tool that created it. Reorder the files if that matters to your CAT tool.
"Isn't valid XML" or "isn't a TMX file". The file may be damaged or be another format. Export it again from your CAT tool. The merge and split tools read TMX only.
Export each one as TMX, drop them all into TMX merge, turn on duplicate removal and download one TMX. Then import it into the tool you use. Files in UTF-8, UTF-16 and Windows-1252 can be mixed.
The first file's. Its source language should match the others; files that differ are listed so you can check them.
No. Each unit is copied exactly as it was, including dates, users and properties. Only duplicates are dropped, and, if you chose one language pair, the other languages.
Use TMX split and split by language pair: each target language becomes its own bilingual TMX with the source language. Or merge with "Only" one pair selected to get a single bilingual file.
500 MB in total for merging and 500 MB for one file when splitting. Very large files depend on your computer's memory.
No. Merging and splitting run in your browser, so the files never reach TraduQ's server.
TraduQ is a marketplace for translators and translation clients, and its file tools are free and run in your browser. After merging, it is worth cleaning the TMX, and before delivering a job, running a QA check.