Find the names added between two releases

You need the same file from both releases, for example the domains file of 28 September 2026 and of 16 October 2026. A purchase includes the release on sale that day.

Why this works

Every file is strictly sorted bytewise, one unique name per line. comm compares two sorted streams line by line; LC_ALL=C makes it compare bytes the same way the files were sorted.

Try it on two small files

printf '%s\n' a.example.com c.example.com > old.txt
printf '%s\n' a.example.com b.example.com c.example.com > new.txt
LC_ALL=C comm -13 old.txt new.txt

It prints b.example.com: the name only in the newer file.

Run it on two releases

In Bash or Zsh:

LC_ALL=C comm -13 <(gzip -dc getdomainlists-2026-09-28-domains.txt.gz) <(gzip -dc getdomainlists-2026-10-16-domains.txt.gz) > added-domains.txt
wc -l added-domains.txt

For these two releases it gives 73,045,002 lines. Every name of the 28 September 2026 release is also in the 16 October 2026 release, so comm -23 (names only in the older file) prints nothing. The subdomains files work the same way, with 444,969,050 more names; they are larger, so that run takes longer.

Without comm

On a system without comm (for example Windows without WSL), the same merge walk in Python 3 streams both files:

import gzip, sys
old, new = (gzip.open(p, "rt") for p in sys.argv[1:3])
o = old.readline()
for n in new:
    while o and o < n:
        o = old.readline()
    if o != n:
        sys.stdout.write(n)

Save it as added.py and run python3 added.py getdomainlists-2026-09-28-domains.txt.gz getdomainlists-2026-10-16-domains.txt.gz > added-domains.txt.

What "added" means

An added name is new to these files, not newly registered: it is in the newer release and not in the earlier one. It is not evidence that the name is registered, resolves or hosts a website today. Each release's figures are in the release notes.

Try the format first with the free 1,000-name samples. Both full files are $9 once, with 30 days to download.

To keep one ending from the added names, pipe them through the TLD filtering guide.