Unix - count unique IP addresses, sort them by most frequent and also sort them by IP when number of repetitions is same

Question

I have a list of IP addresses in file:

  72.204.55.250
  72.204.55.250
  72.204.55.250
  72.204.55.250
  72.204.55.250
  96.41.51.202
  208.115.113.91
  178.137.94.166
  178.137.94.166
  208.115.113.91
  96.41.51.202
  141.8.143.179
  141.8.143.179

Now I am going to sort them and call uniq -c service and I get:

  2  141.8.143.179
  2  178.137.94.166
  2  208.115.113.91
  5  72.204.55.250
  2  96.41.51.202

Now I will sort them by most frequent (sort -rn) but my problem is to also sort them by IP address when the number of repetitions is same in descending order. I've found a sort command for only IP address which works:

sort -rn -t . -k1,1 -k2,2 -k 3,3 -k4,4

but as I mentioned it above, I don't know how to combine it with first column (number of repetitions) to get these expected results:

  5  72.204.55.250
  2  208.115.113.91
  2  178.137.94.166
  2  141.8.143.179
  2  96.41.51.202

How can I achieve that? Thanks in advance for any help.

Solution

If your sort can do a stable sort, e.g. GNU sort with the -s or --stable option, lines with fields unrelated to the sort keys will not be sorted by those fields when there are ties, but will stay in their same relative positions.

$ sort -n -t. -k1,1 -k2,2 -k3,3 -k4,4 | uniq -c | sort -n -r -s
  5 72.204.55.250
  2 96.41.51.202
  2 141.8.143.179
  2 178.137.94.166
  2 208.115.113.91

Comments

[cutrightjm]

Well you can use this to get the quantity sorted in the correct order, but the ip addresses themselves will be in reverse order: sort -rn <file> | uniq -c

[Lkopo]

Exactly and there is the problem - I need them both in the same order.

[Lkopo]

@don_crissti yes it is needed, some IP addresses are not sorted correctly. This is a correct answer for me, just when you want to sort IP addresses in descending order too, just add -r in first sort. Thank you so much!