Bash - Awk排序 - 不排序

Question

awk '{for(i=1; i<=NF; i++) printf("%d ",$i)}' | sort -n

it reads a file like

55 89 33 20

and prints it out normally, not numerically sorted. Why?

Answer 1

sort works on a per-line basis, and printf doesn't append a newline by default, you need to specify it. So use:

awk '{for(i=1; i<=NF; i++) printf("%d\n",$i)}' | sort -n

This will print out your numbers on separate lines, if you want them to be in a single line again then you can pipe it to paste :

awk '{for(i=1; i<=NF; i++) printf("%d\n",$i)}' | sort -n | paste -s -d ' '

You can also just use print instead of printf, this will append the newline by default:

awk '{for(i=1; i<=NF; i++) print $i}' | sort -n

Answer 2

If you want to keep the numbers all on one line but sort them, you can also just do the sorting in awk itself.

$ awk '{split($0,f,FS); n=asort(f); 
     for (i=1; i<=n; ++i) printf("%d ", f[i]); printf "\n"}' <<<'55 89 33 20'
20 33 55 89

But this is starting to get into territory where I reach for Perl instead, just because it's a shorter program to write at the shell prompt:

$ perl -lane 'print join " ",sort @F' <<<'55 89 33 20'
20 33 55 89

Awk is more expressive in many cases, but array operations are not one of them.

Ruby is about as short, if you prefer a language that was cool more recently than Perl :) :

$ ruby -lane 'puts $F.sort.join " "' <<<'55 89 33 20'
20 33 55 89

Even Python, which is not exactly built for one-liners, is less code than Awk in this case (but the same number of parentheses):

$ python -c 'print " ".join(sorted(raw_input().split()))' <<<'55 89 33 20'
20 33 55 89

Answer 3

Setting the file as per your example:

echo "55 89 33 20" > file.txt

The following command will output what you want:

awk '{for(i=1; i<=NF; i++) print($i)}' file.txt | sort -n
20
33
55
89

Note that print will print a new line character each time.
Also, note that the filename should be passed as an argument to awk .

Answer 4

没有awk的解决方案：

cat file | sed 's/ /\n/g' | sort -n | paste -s -d ' '