Hi, I am having a problem with grep while using it in a for loop. I have a file "ab" that looks like : (start file) --comment 1 -- comment 2 hghj jh abc bht nef 987 bht pit --comment 3 890 gyu ppp 456 bht lll (end file) I have to pick out all the lines that contain "bht" and perform some operations on them. I tried using for i in `cat ab |grep "bht"` do echo $i ...do some tasks done But this is not giving me the result I want, since instead of having the output for $i in the form abc bht nef 987 bht pit 456 bht lll I am getting it in the form abc bht nef 987 bht pit 456 bht lll I need the for loop to be there, since there are additional operations to be done in line-to-line basis for all lines containing "bht". Can anyone help? -Anukta

How about this?" for i in `grep " bht " ab` ; do echo "${i}" do tasks ........ done You don't need 'cat' and it looks like the string 'bht' is surrounded by spaces in the ab file.

This works: grep " bht " ab | while read line ; do echo "$line" do something ...... done -jim

Thanks, i tried, but it still doesn't work. Also, the "echo $i" statement that I put in there is just a debugging msg, to see the way output is formatted. I will not need it in the main program. My main problem is how to construct the loop. -Anukta

The 2nd solution worked !!! Thanks a lot. -Anukta

This is how it would work in a for loop: for i in "`grep " bht " ab`" ; do echo "$i" do tasks ........ done -jim

Your original script would have been okay, except that the variable IFS was not set to carriage-return. The default is SPACE so consequently... A B C D E F would produce 6 iterations of the loop. IFS="^M" for i in `cat file` .... would count the newline instead, hence 2 iterations as you required. Best Wishes, Sean