Billion Triples dataset

The complete dataset is composed of a set of smaller datasets. Each download is in one of two formats: (1) WARC or (2) tar.gz. You can read about the WARC format by following this link to the mailing list. The tar.gz format is a tarred and gzipped file containing triples given in the N-Triples syntax.

Note: the Webscope data set comes with a license you need to sign and return before you can download it (see README).

If you are interested in participating in the Billion Triples Track, please join the mailing list.