I would like to precede each line with a number saying how many slashes the line has got.
awk '{ l=$0; gsub("[^/]","",l); print length(l),l }'
This doesn't work bacause l=$0 seems to assign by reference.
How do I dup the string?
Is there a better way to do this with standard UNIX tools? I essentially want to sort a list of filepaths by depth (slash count).
No, awk always do assignment by value, not by reference.
The RHS of variable assignment is an expression, and an expression in awk always return a value. To duplicate a variable, just assigning its value to new variable, you can operate on new variable without affecting the original variable.
In:
$ echo 1 | awk '{l=$0; sub("1","2",l); print l, $0}'
2 1
only value of l was modified, $0 value wasn't changed.
With your requirement in the question, simply do:
awk -F '/' '{print NF-1, $0}' <file
You don't need to do any parsing work, let awk do it all for you before you enter the script body. You only need to extract the information.
Per the manual:
gsub(regexp, replacement [, target])
(if target is omitted, default to $0). gsub() returns the number of substitutions made so in your case you want to replace slashes to get the count:
awk '{l=$0; print gsub("/", ""), l}'
but as Etan Reisner notices, the assignment isn't even needed in this case:
awk '{print gsub("/", "/"), $0}'
l at all. Since you don't modify the line in the first place. Just `{print gsub(/\//, "/"), $0}'.
Use split instead:
split(s, a[, fs ])
Split the string s into array elements a[1], a[2], ..., a[n],
and return n. All elements of the array shall be deleted
before the split is performed. The separation shall be done
with the ERE fs or with the field separator FS if fs is not
given. [. . .]
So, given this input file:
$ cat file
no slashes
one / slash
two / and /
consecutive 3 ///
none
one /
You could do:
$ awk '{ n=split($0,a,"/"); print n-1,$0}' file
0 no slashes
1 one / slash
2 two / and /
3 consecutive 3 ///
0 none
1 one /
FS for something else at the same time.
git ls-files, but I'm solving my problem differently in the end. At least I learned something aboutawk.l, then printingl. Why would you expect it to be unchanged? Presumably, you meant to doawk '{l=$0; gsub("[^/]","",l); print length(l),$0}which does what you want.awk '{l=$0; gsub("[^/]","",l); print length(l),$0}but wroteawk '{l=$0; gsub("[^/]","",l); print length(l),l}. I hate these typos.