UNIX for Dummies Questions & Answers

OK, let's set this up. I have a tab delimited file from excel. In my UNIX shell I have the following lines

Code:

   IFS=`printf "\t"`
   while read LINE_NO SKIP IGNORE_ERRORS OTHER
   do
 ....

This works fine if there is something in every column like this. NOTE, those are tabs, not spaces.

Code:

17      SKIP    FALSE   Endofline

LINE_NO=17
SKIP=SKIP
IGNORE_ERRORS=FALSE
OTHER=Endofline


The problem comes about if one of the columns is a null like this.

Code:

17                FALSE   Endofline

LINE_NO=17
SKIP=FALSE
IGNORE_ERRORS=Endofline
OTHER=

How do I get unix to see that double tab? I am on an AIX box running a ksh shell.

Thanks for your time

Hmm... I had a look at an AIX box (not my usual area) and the ksh man page is a bit of a waste of time, isn't it!

Going back to a HP-UX box, it says:

Code:

    Blank Interpretation
      After parameter and command substitution, the results of substitution
      are scanned for field separator characters (found in IFS), and split
      into distinct arguments where such characters are found.  ksh retains
      explicit null arguments ( or '') but removes implicit null arguments
      (those resulting from parameters that have no values).

So it seems like you're out of luck with that particular method. Have you considered using awk or perl or is there lots of other shell code already written that depends on this part?

After experimenting a little, it seems to behave differently if you use a non-whitespace character as your IFS. So if you run your input data through sed and convert all of the separators to, say, ":", then it does appear to work as you'd expect.

i suggest you change your excel delimiter to something else when you export it.