Is there any better solution to reverse uniq count

Question

I want to reverse the uniq -c output from:

  1 hi
  2 test
  3 try

to:

hi
test  
test  
try  
try  
try

My solution now is to use a loop:

while read a b; do yes $b |head -n $a ;done <test.txt

I wonder if there are any simpler commands to achieve that?

I think your solution is awesome. Very creative use of `yes`. — Hubert Grzeskowiak, Jan 02 '18 at 05:02

score 3 · Accepted Answer · answered Jan 02 '18 at 19:27

3

another awk

awk '{while ($1--) print $2}' file

answered Jan 02 '18 at 19:27

thanasisp

5,855
3
14
31

simplest solution by now, may i ask about '($1--)'? – once Jan 03 '18 at 01:47
1

this condition returns `$1` and decrease it by one. Similar to other languages that have the `++`,`--` operands, reduction happens after every evaluation. So it counts downwards from `$1` to zero, for every line. – thanasisp Jan 03 '18 at 11:03

score 2 · Answer 2 · answered Jan 02 '18 at 05:35

Here's another solution

echo "\
1 hi
2 test
3 try" | awk '{for(i=1;i<=$1;i++){print($2)}}'

output

hi
test
test
try
try
try

This will work the same way, with

awk '{for(i=1;i<=$1;i++){print($2)}}' uniq_counts.txt

The core of the script, if of course

 { for (i=1;i<=$1;i++) { print $2 }

where awk has parsed the input into 2 fields, $1 being the number (1,2,3) and $2 is the value, (hi,test,try).

The condition i<=$1 tests that the counter i has not incremented beyond the count supplied in field $1, and the print $2 prints the value each time that i<=$1 condition is true.

IHTH

score 2 · Answer 3 · answered Jan 02 '18 at 05:44

2

You don't need awk or any other command, you can do it entirely in bash

while read n s
do
    for ((i=0; i<n; i++))
    do
        echo $s;
    done
done < test.txt

answered Jan 02 '18 at 05:44

Diego Torres Milano

65,697
9
111
134

1

Quote the expansion and you might also want to use `read -r ...`. – PesaThe Jan 02 '18 at 10:04

Jdamian · Answer 4 · 2018-01-02T18:57:52.700

Here my solution uses the bash brace expansion and the printf internal command.

while read a b
do
     eval printf "'%.${#b}s\n'" "'$b'"{1..$a}
done <test.txt

The following simple example

printf '%s\n' test{1..2}

prints two lines which contain the string test followed by a number:

test1
test2

but we can specify the exact number of characters to print by using the precision field of the printf command:

printf '%.4s\n' test{1..2}

to display:

test
test

The length of the characters to print is given by the length of the text to print (${#b}).

Finally the eval command must be used in other to use variables in the brace expansion.

Is there any better solution to reverse uniq count

4 Answers4