UNIX for Dummies Questions & Answers

HP-UX

I have a fixed length file like this

9921190625797AE2560 20091001US20091001@@NEWSITE @@ 20091013001X X 01NEW00DNA00007081 @@ SPRINGFIELD @@ @@ NJ US 001 USX

I want to switch the columns and replace certain values into another file. For example, the output file will have the columns in the following order from input file

chars 1-13
chars 81-98 (if input is spaces replace with current date)
chars 14-23
chars 24-31 (if input is spaces replace with "00010101")
....
....
and so forth

I can switch the columns using the below awk script but not sure about replacing the values.

Code:

tmp.awk

{print \
substr($0,1,13)  \
substr($0,81,8)  \
substr($0,14,10) \
substr($0, 24, 8) \
substr($0,32,2)}

cat input_file | awk -f tmp.awk

To replace space columns from 24 to 31:

Code:

mTo='00010101'
sed "s/\(.\{23\}\)[ ]\{8\}\(.*\)/\1${mTo}\2/" Inp_File

You can run it before or after your awk.

Thanks a lot

How to check if column has a certain value? For example, I want to check if columns from 50 to 60 has '@@' then replace with spaces. Can I perform multiple find and replace with one call to sed or they need to be done separately?