# frequency count using shell **URL:** \ **Category:** Shell Programming and Scripting\ **Created:** [October 9, 2012, 4:51pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968 "2012-10-09T16:51:46Z")\ **Posts on this page:** 8\ **Page:** 1
**Author:** ![xshang](https://community.unix.com/letter_avatar/xshang/32/6_22734026db12803ebb3e059622ac3534.png) [@xshang](https://community.unix.com/u/xshang)\ **Post date:** [October 9, 2012, 4:51pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/1 "2012-10-09T16:51:46Z")
Hello everyone, please consider the following lines of a matrix ```auto [574,] 59 32 [575,] 59 32 [576,] 59 32 [577,] 59 32 [578,] 59 32 [579,] 59 32 [580,] 59 32 [581,] 60 32 [582,] 60 33 [583,] 60 33 [584,] 60 33 [585,] 60 33 [586,] 60 33 [587,] 60 33 [588,] 60 33 [589,] 60 33 [590,] 60 33 [591,] 61 33 [592,] 61 33 [593,] 61 33 [594,] 61 33 [595,] 61 33 [596,] 61 33 [597,] 61 33 [598,] 61 33 [599,] 61 33 [600,] 61 33 [601,] 62 34 ``` Is is possible to count the percent frequency of each distinct field in $2? Just like this: ```auto 59 25.00% 60 35.70% 61 35.70% 62 3.57% ``` ---
**Author:** ![Corona688](https://community.unix.com/letter_avatar/corona688/32/6_22734026db12803ebb3e059622ac3534.png) [@Corona688](https://community.unix.com/u/Corona688)\ **Post date:** [October 9, 2012, 5:01pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/2 "2012-10-09T17:01:50Z")
```nohighlight awk '{ A[$2]++ } END { for(X in A) printf("%s\t%s\n", X, (A[X]*100)/NR) }' inputfile ``` ---
**Author:** ![fastlane3000](https://community.unix.com/letter_avatar/fastlane3000/32/6_22734026db12803ebb3e059622ac3534.png) [@fastlane3000](https://community.unix.com/u/fastlane3000)\ **Post date:** [October 9, 2012, 5:25pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/3 "2012-10-09T17:25:33Z")
in case you need to add "%" and use float ```nohighlight awk '{ A[$2]++ } END { for(X in A) printf("%d\t%.2f%\n", X, (A[X]*100)/NR) }' inputfile ``` ---
**Author:** ![rdrtx1](https://community.unix.com/user_avatar/community.unix.com/rdrtx1/32/2721_2.png) [@rdrtx1](https://community.unix.com/u/rdrtx1)\ **Post date:** [October 9, 2012, 5:30pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/4 "2012-10-09T17:30:55Z")
Same solution, added format: ```nohighlight awk '{ A[$2]++ } END { for(X in A) printf("%s\t%5.2f%%\n", X, (A[X]*100)/NR) }' inputfile | sort -n ``` ---
**Author:** ![xshang](https://community.unix.com/letter_avatar/xshang/32/6_22734026db12803ebb3e059622ac3534.png) [@xshang](https://community.unix.com/u/xshang)\ **Post date:** [October 9, 2012, 5:37pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/5 "2012-10-09T17:37:17Z")
> [@fastlane3000](#): > > in case you need to add "%" and use float > > ```plaintext > awk '{ A[$2]++ } END { for(X in A) printf("%d\t%.2f%\n", X, (A[X]*100)/NR) }' inputfile > > ``` It showed some message like this, how should I adjust the code? I'm not familiar with printf, Thank you! ```html awk: weird printf conversion % input record number 61124, file HB143-0W-A4.txt source line number 1 awk: not enough args in printf(%d %.2f% ) input record number 61124, file HB143-0W-A4.txt source line number 1 ``` ---
**Author:** ![rdrtx1](https://community.unix.com/user_avatar/community.unix.com/rdrtx1/32/2721_2.png) [@rdrtx1](https://community.unix.com/u/rdrtx1)\ **Post date:** [October 9, 2012, 5:44pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/6 "2012-10-09T17:44:28Z")
```nohighlight awk '{ A[$2]++ } END { for(X in A) printf("%d\t%.2f%%\n", X, (A[X]*100)/NR) }' inputfile ``` ---
**Author:** ![fastlane3000](https://community.unix.com/letter_avatar/fastlane3000/32/6_22734026db12803ebb3e059622ac3534.png) [@fastlane3000](https://community.unix.com/u/fastlane3000)\ **Post date:** [October 9, 2012, 5:45pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/7 "2012-10-09T17:45:53Z")
maybe you're not using the same awk than i have (i've got a gnu version called gawk) just try "[rdrtx1](http://www.unix.com/members/302126107.html)"'s solution it's more complete with a "sort" ```plaintext awk '{ A[$2]++ } END { for(X in A) printf("%s\t%3.2f%\n", X, (A[X]*100)/NR) }' inputfile | sort -n ``` i rather use 3 it's enough because we can't have more than 100% ---
**Author:** ![alister](https://community.unix.com/letter_avatar/alister/32/6_22734026db12803ebb3e059622ac3534.png) [@alister](https://community.unix.com/u/alister)\ **Post date:** [October 9, 2012, 5:58pm UTC](https://community.unix.com/t/frequency-count-using-shell/317968/8 "2012-10-09T17:58:57Z")
Nit: %s makes more sense than %d since $2 (aka X) is treated as a string in every expression in which it appears. Regards, Alister