easy way to change the uniq -c output?

Question

I have a simple file like this:

Term1 column2 column3
Term2 column2 column3
Term3 column2 column3
Term2 column2 column3
Term1 column2 column3
Term2 column2 column3

If I sort on the first column and get a count for the terms:

cut -f1 -d ' ' file | sort | uniq -c

The results shows me this:

    2 Term1
    3 Term2
    1 Term3

But I would rather see this:

Term1 2
Term2 3
Term3 1

Is there an easy way to "change" the uniq output to do this or would i still need to pipe the output to sed or awk to get this?

This seems a simple question so my apologies if this was asked before.

Please use code tags for Input samples rather than quote tags. — RavinderSingh13, Oct 17 '17 at 10:29

Rahul Verma · Accepted Answer · 2017-10-17T11:13:20.720

5

Uniq doesn't give the option to rearrange output.

You can do this using awk :

cut -f1 -d ' ' file | sort | uniq -c | awk '{print $2, $1}'

Term1 2
Term2 3
Term3 1

Or using awk and then sort (recommended)

$awk '{a[$1]++} END{for (i in a) print i,a[i] }' file | sort -k1
Term1 2
Term2 3
Term3 1

edited Oct 17 '17 at 11:13

answered Oct 17 '17 at 10:06

Rahul Verma

awk one-liner won't guarantee the sorted output. `(i in a)` – Kent Oct 17 '17 at 11:07
`-k1` is redundant. – karakfa Oct 17 '17 at 16:36
@karakfa: Can you elaborate ? – Rahul Verma Oct 17 '17 at 16:40
`sort -k1` is the default behavior of `sort`, no need to explicitly specify. Unless you meant `-k1,1` – karakfa Oct 17 '17 at 17:15
@karakfa: Oh okay. – Rahul Verma Oct 17 '17 at 18:39

score 2 · Answer 2 · answered Oct 17 '17 at 11:36

2

The shortest one with GNU datamash tool:

datamash -Ws -g1 count 1 <file

The output:

Term1   2
Term2   3
Term3   1

answered Oct 17 '17 at 11:36

RomanPerekhrest

2 Answers2