Question

I search a tool(chain) (sed/bash/…) in Linux, to achieve the following: Go through a text file (here: XML) find same values, check these values for some rule and correct these values by adding an offset.

A example would be:

<!--Origin.txt-->
<Tree>
  <Node foo="600" bar="50" />
  <Node foo="-300" bar="600" />
</Tree>

I now wan’t to ignore bar completely, but check if e.g. foo is smaller than zero. If it is, add 1000 to foo. If not, leave it.

<!--Output.txt-->
<Tree>
  <Node foo="600" bar="50" />
  <Node foo="700" bar="600" />
</Tree>

I could write a small python/java program, but I bet there is a faster bash/sed/… script to do this 😉

Answer 1

open your file with vim, then type:

:%s/foo="\zs\(-[0-9]\+\)"\ze/\=submatch(1)+1000/g

it should do the job 4 u.

EDIT

awk oneliner works for your new requirement.

 awk '{for(i=1;i<=NF;i++)
if($i~/foo="/){ split($i,x,"\""); x[2]=x[2]<100?x[2]+1000:x[2]; $i="foo=\""x[2]"\""; }}1' yourFile

in above oneliner, the criteria is the foo attribute value <100. you can add criterias as you wish. like (x[2]>100 && x[2]<700)