I have a huge text file with something like that:
igflex01; igflexdev01; igdbstg01;
igldapint01; igjetdev01; igsql01;
igesxstg01; igdbint01; igdmstg01;
igdsdev01; igdmmmo01; igmsgint02;
and so on...
I would like to extract all text between semicolons when the string 'int' is present. Is there a way to do that?
Since you asked for sed / grep, here it is:
> cat file
igflex01; igflexdev01; igdbstg01;
igldapint01; igjetdev01; igsql01;
igesxstg01; igdbint01; igdmstg01;
igdsdev01; igdmmmo01; igmsgint02;
> sed -nr 's/\s*;\s*/\n/gp' file | grep int
igldapint01
igdbint01
igmsgint02
Or get it with only a single sed call...
> sed -nr -e 's/\s*;\s*/\n/g;/int/P' file
igldapint01
igdbint01
igmsgint02
Used the "tr" command to replace the semicolons with newlines and then run grep (which is line-oriented).
$ cat input.txt | tr ';' '\n' | grep int
igldapint01
igdbint01
igmsgint02
One way is to cheat and convert the ; into newlines and then you can just grep
eg
tr ';' '\012' | grep int
eg
$ cat x
igflex01; igflexdev01; igdbstg01;
igldapint01; igjetdev01; igsql01;
igesxstg01; igdbint01; igdmstg01;
igdsdev01; igdmmmo01; igmsgint02;
$ cat x | tr ';' '\012' | grep int
igldapint01
igdbint01
igmsgint02
Note that it has the spaces in some cases because the space is in between the ;s. Also note that the first word of a line (which isn't strictly between ; on a line, but is between them in the data) is also returned.
% echo 'aint;int;cint' | perl -nle 'print for grep /int/, split /\s*;\s*/'
aint
int
cint
%
Or just split ";" to preserve any whitespace between.
$ cat abc.txt
igflex01; igflexdev01; igdbstg01;
igldapint01; igjetdev01; igsql01;
igesxstg01; igdbint01; igdmstg01;
igdsdev01; igdmmmo01; igmsgint02;
$ grep -o '[^;]*int[^;]*' abc.txt
igldapint01
igdbint01
igmsgint02
If spaces are not desired:
$ grep -o '[^ ;]*int[^ ;]*' abc.txt
igldapint01
igdbint01
igmsgint02
