The crawl archive for February 2019 is now available! It contains 2.9 billion web pages or 225 TiB of uncompressed content, crawled between February 15th and 24th.
The February crawl contains page captures of 750 million URLs not contained in any crawl archive before. New URLs are sampled based on the host and domain ranks (harmonic centrality) published as part of the Nov/Dec/Jan 2018/2019 webgraph data set from the following sources:
- sitemaps, RSS and Atom feeds
- a breadth-first side crawl within a maximum of 5 links (“hops”) away from the homepages of the top 50 million hosts and domains
- a random sample of outlinks taken from WAT files of the January crawl
The number of sampled URLs per domain depends on the domain’s harmonic centrality rank in the webgraph data set – higher ranking domain are allowed to “contribute” more URLs.
The way our crawler handles politeness limits per host and/or pay-level domain has been improved:
First, limits are now configurable and are based on the harmonic centrality rank of a domain.
Second, we now also put a limit on the number of hosts/subdomains per domain. This limit is also based on the domain rank and ranges from 500,000 subdomains for top-ranking domains (think of blogspot.com) to less than 100 for low-ranking domains. While the the number of hosts covered in the February crawl dropped to 50 millions from 60 millions in January, we see a positive impact on the total amount of pages crawled for large domains. Technically, every host requires a DNS lookup and a robots.txt fetch even if only a single page is fetched from this host and the performance of the crawler improves if resources are focused on few 100,000 subdomains and not spread over millions of hosts. We also hope that a limit on the number of hosts per domain makes the crawler more robust against link spam. The set of sampled subdomains for large domains will vary from month to month to guarantee a good overall coverage if multiple monthly crawls are combined.
Archive Location and Download
The February crawl archive is located in the commoncrawl bucket at crawl-data/CC-MAIN-2019-09/.
To assist with exploring and using the dataset, we provide gzipped files which list all segments, WARC, WAT and WET files.
By simply adding either s3://commoncrawl/ or https://commoncrawl.s3.amazonaws.com/ to each line, you end up with the S3 and HTTP paths respectively.
File List #Files Total Size
Segments CC-MAIN-2019-09/segment.paths.gz 100 WARC files CC-MAIN-2019-09/warc.paths.gz 64000 59.86
WAT files CC-MAIN-2019-09/wat.paths.gz 64000 18.23
WET files CC-MAIN-2019-09/wet.paths.gz 64000 7.62
Robots.txt files CC-MAIN-2019-09/robotstxt.paths.gz 64000 0.17
Non-200 responses files CC-MAIN-2019-09/non200responses.paths.gz 64000 1.79
URL index files CC-MAIN-2019-09/cc-index.paths.gz 302 0.21
Please donate to Common Crawl if you appreciate our free datasets! We’re also seeking corporate sponsors to partner with Common Crawl for our non-profit work in open data. Please contact [email protected] for sponsorship information.