Shell Programming and Scripting

Code:

sed '/abc/s/\(.*[ \t]*\)\(abc.[^ \t]*\)\(.*\)/\2/' file

Thanks for the reply Radoulov (and others). As I could not access my documents yesterday I have discovered a slight problem, and although the solution partially works it doesn't do as required, which is partly my fault. The ID number is actually of the format:

abcd-p-ssa-a322f-s-4312

with the hyphens, numbers, and letters in that order. Currently the solution will only print upto the abc and not everything else. However, to add to my misery, there are also other types of ID of the format:

abcf-p-ssa-a322f-s-4312 (the abc here still remains)
def-abcf-p-ssa-a322f-s-4312 (addition of def- at the start of the two above)
def-abcd-p-ssa-a322f-s-4312

The first few sets of lettters always remain the same (i.e. abc-p-ssa , abcf-p-ssa , def-abcf-p-ssa , def-abcd-p-ssa) . Sorry for the inconvienence caused (I probably spelt that wrong!). Thank you for your time.

---------- Post updated at 10:45 AM ---------- Previous update was at 09:29 AM ----------

Hey, I tried this out but it cuts out the last line of my file for some reason. Also if you could, my requirements have now changed slightly they are written above. Thanks

Code:

perl -nle'
   /abc/ and print join " ", /abc[\w-]+/g
     or print
    ' infile

Thats world class mate! But if I also wanted to add the option to search for lines with the def instead of abc in addition to the abc, how would I do that? Is there some sort of OR operator?

Code:

perl -nle'
  /(?:abc|def)/ and print join " ", /(?:abc|def)[\w-]+/g      
    or print     
    ' infile

Brilliant! Thanks mate, really appreciate the help!

sorry to bring this up again, but since this answer theres been a few changes in my input file,mainly the addition of a new set of numbers

Code:

abcf-p-ssa-a322f-s-4312 
def-abcf-p-ssa-a322f-s-4312 
def-abcd-p-ssa-a322f-s-4312
#below is the new addition
ghi/123/xxx/sss/xa2
ghi/3d/cksdi/kff/23 def-abcd-p-ssa-a322f-s-4312

the ghi always remains at the start, so the problem is that the code you gave doesn't work because of the brackets in this new set, as I tried:

Code:

perl -nle'
  /(?:abc|def)/ and print join " ", /(?:abc|def)[\w-]+/g      
    or print     
    ' infile

but that just ignored anything with brackets! I also removed the '+' sign after [\w-] but that gave me the ghi, but not the rest of the number! Also as in teh last line, some of these may be on the same line, but your code already seems to work for that anyway (excpet when its got '/' in the number, so I need to keep that functionality)

Many thanks