The LDBC SNB datasets are available through Cloudflare R2.

If you experience issues when downloading large files from Cloudflare R2, try this “strenghtened” wget command that will insist on retrying in case of intermittent failures:

wget --continue \
     --tries=0 \
     --waitretry=30 \
     --retry-connrefused \
     --retry-on-http-error=429,500,502,503,504 \
     --read-timeout=30 \
     --timeout=60 \
     "https://datasets.ldbcouncil.org/..."

SNB Interactive v1

Initial datasets

format \ sf 0.1 0.3 1 3 10
CsvBasic & LongDateFormatter
CsvComposite & LongDateFormatter
CsvMergeForeign & LongDateFormatter
CsvCompositeMergeForeign & LongDateFormatter
CsvBasic & StringDateFormatter
CsvComposite & StringDateFormatter
CsvMergeForeign & StringDateFormatter
CsvCompositeMergeForeign & StringDateFormatter
Turtle
format \ sf 30 100 300 1000
CsvBasic & LongDateFormatter
CsvComposite & LongDateFormatter
CsvMergeForeign & LongDateFormatter
CsvCompositeMergeForeign & LongDateFormatter
CsvBasic & StringDateFormatter
CsvComposite & StringDateFormatter
CsvMergeForeign & StringDateFormatter
CsvCompositeMergeForeign & StringDateFormatter
Turtle

Update streams

#parts \ sf 0.1 0.3 1 3 10 30 100 300 1000
1
2
4
8
16
24
32
48
64
96
128
192
256
384
512
768
1024

Parameters

SF3000

The SNB Interactive SF3000 data set, update streams, and parameters are available in a single format:

SNB Business Intelligence

Compressed CSVs in the composite-merged-fk format

Checksums: bi-composite-merged-fk-md5sums.tar.zst

Compressed CSVs in the composite-projected-fk format

Checksums: bi-composite-projected-fk-md5sums.tar.zst

Compressed CSVs in the composite-projected-fk CSV format with quotes and without headers

Checksums: bi-composite-projected-fk-with-quotes-without-headers-md5sums.tar.zst

Raw (up to SF30)

Checksums: bi-raw-md5sums.tar.zst

Factor tables

Checksums: bi-factors-md5sums.tar.zst

The SF30k factors were generated with a newer version of the data generator.