for id in `cat ids|sort -rn|uniq`; do echo `grep -w $id ids|wc -l` :$id; done|sort -nr
Refactorings
No refactoring yet !
Nick
November 1, 2008, November 01, 2008 16:57, permalink
1 rating.
Login to rate!
I'd write a small script for this. I've been using Ruby lately, so here's one solution:
filename = 'process.log'
ids = {}
File.readlines(filename).each do |line|
id = line.match(/\d+/)[0]
ids[id] ? ids[id] += 1 : ids[id] = 1
end
ids.sort {|a,b| a[1] <=> b[1]}.reverse.each do |pair|
puts "#{pair[1]} :#{pair[0]}"
end
I was processing logs like this:
$ tail -1000000 process.log|grep 'PROCESSING N'|cut -d: -f2 > ids
$ cat ids
1
2
3
4
6
8
26
44
54
55
56
83
87
93
95
121
129
139
143
159
161
165
56
83
159
56
83
159
169
56
83
159
56
83
159
173
56
83
159
56
83
159
175
56
83
159
Then to get stats:
$for id in `cat ids|sort -rn|uniq`; do echo `grep -w $id ids|wc -l` :$id; done|sort -nr
8 :83
8 :56
8 :159
1 :95
1 :93
1 :87
1 :8
1 :6
1 :55
1 :54
1 :44
1 :4
1 :3
1 :26
1 :2
1 :175
1 :173
1 :169
1 :165
1 :161
1 :143
1 :139
1 :129
1 :121
1 :1
Can this be simplified somehow?