66

I'm looking for a list of German words in a digital format that is simple and easy to parse by a computer. It should encompass almost all words, but it doesn't matter whether inflections are included. Acronyms and proper names are not interesting, but those are easy to filter.

The Wiktionary dump does not fulfill the criterion of being simple to parse, since I'd have to check each article whether it is a German word or not, nor the all-encompassing criterion.

Bonus points if lexical class is included and the license is open.

Improve this question
0

9 Answers 9

Reset to default
38

Here are some places where you can download such lists:

Improve this answer
2
  • 1
    If only that list had definitions :D By the way, is it free to use?^
    – Alenanno
    Commented May 28, 2011 at 9:44
  • Be aware, that a lot of the entries in there are English words. It might be somewhat helpful if you want an idea of the English words Germans like to use in practice e.g. "sales" but not "salesman" + a bunch of swearwords.
    – Adamantish
    Commented Mar 28, 2019 at 13:11
21

In most recent Linux distributions there are two files:

/usr/share/dict/ogerman
/usr/share/dict/ngerman

These contain a list of line-separated German words. ogerman is for the old-spelling and ngerman is for reformed spelling. On my system, ogerman contains about 76000 words, while ngerman has about 330000 words.

It may be needed to install a package containing those files, and the path may be different either. Under debian, those are released under the GPL license, the package names are wogerman and wngerman.

Improve this answer
2
  • I just wondered, as dpkg -S ngerman only gave me texlive packages. But I guess it only uses installed packages.
    – mbx
    Commented May 28, 2011 at 14:56
  • 2
    @mbx There is a w in fornt of ngerman. You could use apt-file to find that package. AFAIK, wogerman is preinstalled if you have full German language support. BTW, using these packages is great too, since you can easily put them as a dependency of your project instead of shipping them.
    – FUZxxl
    Commented May 28, 2011 at 15:07
16

I realise this is a bit late, but I thought I'd share this here in the chance that it helps someone else coming here from a search engine.

There is a large text file available from dict.cc. It is quite comprehensive and includes whether the word is m/f/n/pl, and also the English translation.

http://www1.dict.cc/translation_file_request.php?l=e

Other languages and directions are also available.

Improve this answer
1
  • 3
    Alas, the extremely restrictive license on dict.cc disallows its use in all modern German learning software. Even personal use on any online memorization site is prohibited.
    – samkass
    Commented Dec 26, 2019 at 13:16
7

Google's NGramViewer also offers the raw datasets for download. The datasets also contain the number of occurences of the word (and for combinations of up to five words) in any given year. This may be useful for statistics on how differend word usages evolved; it's used all the time on EL&U.

Improve this answer
6

You can download 431000 word forms in XML format, parsed from the morphy tool. The licence is " Creative Commons Namensnennung - Weitergabe unter gleichen Bedingungen 4.0 Unported".

Improve this answer
5

I found a postgresql-database-dict on my Ubuntu system

/var/cache/postgresql/dicts/de_de.dict

The size is about 1Mb, it isn't compressed, and the content around Ab-Ac looks like this:

Abzug/STpm
Abzugs/hij
Abzweig/hij
Abzweigen/SJm
Accessoire/Sm
Account/Sm
Accounting/Sm
Aceton/Sm
Achilles/ijm
Achim/S
Achs/hij
Achse/Nm

Then there is 

 /usr/share/hunspell/de_DE.dic

nearly of same size and format as the postgres file.

Additionally, there are, beside (n|o)german-files, mentioned by FUZxxl, aspell- and ispell-files, which are compressed somehow, but I don't know how. Using gunzip and word-list-compress -d didn't work, as mentioned in the manpage. 

/var/lib/aspell/de-alt-common.rws
/usr/share/aspell/de-alt-common.cwl.gz
/usr/share/aspell/de-common.cwl.gz
Improve this answer
4

http://www.wortschatz.uni-leipzig.de/html/wliste.html offers lists of the most used words up to 10.000 words. These lists are plain text, they should be easy to parse.

Improve this answer
3
  • The format looks great. Dudens Universalwörterbuch lists half a million words, though, so 10,000 doesn't come close to all words.
    – Tim
    Commented May 28, 2011 at 7:18
  • 2
    Bei einem Test bekomme ich heute "File not found". Vielleicht mal testen, ob das andauert - die Liste womöglich umgezogen ist.
    – user unknown
    Commented Mar 29, 2017 at 2:16
  • The link is dead. I didn't find the 10,000 words list, but here are other downloads from Uni Leipzig: wortschatz.uni-leipzig.de/de/download
    – Hubert Schölnast
    Commented Feb 16, 2023 at 6:18
3

The Kaikki project has a parsed (JSON) version of the German entries of English Wiktionary, including inflections, definitions and categories.

It also has a downloadable version of the German Wiktionary with German glosses.

Improve this answer
-1

I came across this question when googling for "Parsable list of german words"

The following repo has a .csv of german words with their gender:

kalbsschnitzel  n
luftmatratze    f
schraubenwelle  f
mitschnitt      m

https://github.com/Pimverleg/German-words-with-gender

Improve this answer
3
  • 3
    This list assigns a gender to words of each part of speech, which is nonsense. In fact, only nouns have genders. The list assigns genders to pronouns, verbs, adverbs, numerals, etc., which makes absolutely no sense: The word "ich" is listed as a neuter, "was" as a masculine, "ist" = neuter "habe" = feminine, "jetzt" = neuter, "zwei" = feminine. This all is simply wrong! The list also completely ignores the fact that German nouns always begin with a capital letter. The noun "Leben" is a neuter noun but is missing from the list. Instead, the verb "leben", which in fact has no ...
    – Hubert Schölnast
    Commented Feb 16, 2023 at 6:59
  • 3
    ... gender, is there in the list as a neuter. Nouns with more than one gender are listed only once: "Event" is listed only as a masculine (der Event). The neuter (das Event) is not present, although it is used as frequently as the masculine. The same is true for "Filter," "Gummi," and all other multi-gendered words. Some nouns are present in the list only in the plural form, or only as part of compound words: "Kraken" is present in the list, but the singular "Krake" is missing, and the word "Naturjoghurt" is present, but "Joghurt" is not.
    – Hubert Schölnast
    Commented Feb 16, 2023 at 6:59
  • @HubertSchölnast Thank you for your feedback, the list is a work in progress. I hope to update the repo soon using your feedback
    – Mister Verleg
    Commented Feb 23, 2023 at 12:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.