Shell Programming and Scripting

As the OP is running ksh, no need to set bash specific options, the same loop works as is:

Code:

while read line
do
        echo $line | IFS="," read dbt_acct val_dt crncy AMT dbt_GOP
        print $AMT
done <abc.txt

Hi Jlliagre,
I have not said otherwise, I just made a remark about the response RudiC about bash.

Regards.

Adding some debug code to your original script

Code:

#!/bin/ksh
debugfile=debug.txt
while read line
do
  echo "$line" |
   awk -F"," '{print $1" "$2" "$3" "$4" "$5}' |
   tee -a $debugfile |
   read dbt_acct val_dt crncy AMT dbt_GOP
  print "$AMT"
done <abc.txt

Now look at the debug.txt file, and you should see the problem!

Code:

123 ra point  there
232 ba points home pheer

Hi rahulsk,
In case it isn't obvious from what others have already said, your code didn't work because your awk script produces output (when there are one or more empty fields in the inpu) that your second read statement interprets to be part of a field separator instead of an empty field.

Furthermore, your use of echo (with an unquoted variable expansion) would convert any sequence of one or more <space> and <tab> characters to a single <space> character. (This doesn't matter with the supplied sample data, but if your input was something like:

Code:

123,ra,point	x y,,there
232,ba,points	a b	c d	e f,home,pheer

it could make a huge difference. And, with input like the above, the echo and awk pipeline converting <comma>s to <space>s would further exacerbate the problem changing <space> and <tab> characters in your input data into field separators and may add lots of words to the contents of the last data field.)

Using a single read statement with a <comma> as the field separator instead of a read statement, an echo statement, an awk command, and another read statement gets rid of two unnecessary writes of your data, gets rid of two unnecessary reads of your data, gets rid of the invocation of two unneeded shell built-in utilities, and gets rid of one unneeded external utility invocation as well as getting rid of several places in the pipeline where data can be corrupted.

A little trick to work with awk|read:

Code:

$ while read line; do echo $line |awk -F"," '{gsub(/,,/,",\\\000,");print $1" "$2" "$3" "$4" "$5}'|read dbt_acct val_dt crncy AMT dbt_GOP; echo $AMT; done <abc.txt

home

But, it is not the best solution, just to show it is possible...

Regards.