NSRL Unique File Corpus
NIST's National Software Reference Library makes available a Unique File Corpus of each file that it processes. The tab delimited file lists the SHA1 hash value and byte count of each file. It is available for download here.
If you're conducting electronic discovery and find unknown files and want to determine if they may contain data unique to a business's network consulting the corpus is one way to 'de-NIST' the data.
The full NSLR Reference Data Set contains far more information for each file, including the CRC32 checksum: the MD5 hash value; manufacturer, operating system, and product codes, names, and versions; language; application type; and the file name. Whereas the full set downloads as an .iso file, the Unique File Corpus downloads as a single file named, 'CorpIdMetadata.tab' in a zip file.
The tab format can be handled by MS Access. I was able to successfully import the first 1,048,576 rows into an Excel spreadsheet tonight. Take a look: