mjhjmtu

I have a string that looks something like
/foo/bar/baz59_ 5stuff.thing

I would like to increase the last number (5 in the example) by one, if it's greater than another variable. A 9 would be increased to 10. Note that the last number could be multiple digits also; also that "stuff.thing" could be anything; other than a number; so it can't be hard coded.

The above example would result in /foo/bar/baz59_ 6stuff.thing

I've found multiple questions (and answers) that would extract the last number from the string, and obviously that could then be used in a comparison.
The issue I'm having is how to ensure that when I do the replace, I only replace the last number (since obviously I can't just replace "5" for "6"). Can anyone make any suggestions?

awk/sed/bash/grep are all viable.

Updated Answer

Thanks to @EdMorton for pointing out the further requirement of the number exceeding a threshold. That can be done like this:

perl -spe 's/(d+)(?!.*d+)/$1>$thresh? $1+1 : $1/e' <<< "abc123_456.txt" -- -thresh=500

Original Answer

You can evaluate/calculate a replacement with /e in Perl regexes. Here I just add 1 to the captured string of digits but you can do more complicated stuff:

perl -pe 's/(d+)(?!.*d+)/$1+1/e' <<< "abc123_456.txt"
abc123_457.txt

The (?!.*d+) is (hopefully) a negative look-ahead for any more digits.

The $1 represents any sequence of digits captured in the capture group (d+).

Note that this would need modification to handle decimal numbers, negative numbers and scientific notation - but that is possible.

Using bash regular expression matching:

$ f="/foo/bar/baz59_ 99stuff.thing"
$ [[ $f =~ ([0-9]+)([^0-9]+)$ ]]

OK, what do we have now?

$ declare -p BASH_REMATCH
declare -ar BASH_REMATCH=([0]="99stuff.thing" [1]="99" [2]="stuff.thing")

So we can construct the new filename

if [[ $f =~ ([0-9]+)([^0-9]+)$ ]]; then
 prefix=$f%$BASH_REMATCH[0] # remove "99stuff.thing" from $f
 number=$(( 10#$BASH_REMATCH[1] + 1 )) # use "10#" to force base10
 new=$prefix$number$BASH_REMATCH[2]
 echo $new
fi
# => /foo/bar/baz59_ 100stuff.thing

With GNU awk for the 3rd arg to match():

$ awk -v t=3 'match($0,/(.*)([0-9]+)([^0-9]*)$/,a) && a[2]>ta[2]++; $0=a[1] a[2] a[3] 1' file
/foo/bar/baz59_ 6stuff.thing

Just set t to whatever your threshold value is for incrementing, e.g.:

$ awk -v t=7 'match($0,/(.*)([0-9]+)([^0-9]*)$/,a) && a[2]>ta[2]++; $0=a[1] a[2] a[3] 1' file
/foo/bar/baz59_ 5stuff.thing

if it's greater than a script argument.

If I get it correctly(I am assuming you are passing an argument through a script and if its value is greater than string's 2nd field digit then increase 1 into that 2nd field's digit), could you please try following once.

cat script.ksh
value=$1
echo "/foo/bar/baz59_ 5stuff.thing" | 
awk -v arg="$value" '
 match($2,/[0-9]+/)
 val=substr($2,RSTART,RLENGTH)
 val=val<arg?val+1:val
 $2=val substr($2,RSTART+RLENGTH)

1'

Here is an example when I run script.ksh it gives following output.

/script.ksh 7
/foo/bar/baz59_ 6stuff.thing

Here is a shorter gnu awk approach:

cat incr.awk

 n = split($0, a, /[0-9]+/, b)
 for(i=1; i<n; i++)
 s = s a[i] b[i] + (b[i] < max && i == n-1 ? 1 : 0)
 print s a[i]

Then use it as:

awk -v max=80 -f incr.awk <<< '/foo/bar/baz59_ 5stuff.thing'
/foo/bar/baz59_ 6stuff.thing

awk -v max=80 -f incr.awk <<< '/foo/bar/baz59_ 79stuff.thing'
/foo/bar/baz59_ 80stuff.thing


awk -v max=80 -f incr.awk <<< '/foo/bar/baz59_ 90stuff.thing'
/foo/bar/baz59_ 90stuff.thing

awk -v max=80 -f incr.awk <<< '/foo/bar/baz59_ 80stuff.thing'
/foo/bar/baz59_ 80stuff.thing

awk -v max=80 -f incr.awk <<< '/foo/bar/stuff.thing'
/foo/bar/stuff.thing