Okay - my mistake - it appears cut was improved on January 9th 2003 but I didn't remember it had that limitation before.

0

Never-the-less, I am always learning something new from you guys, and this time it was Wufgelazabarak!

0

Thanks so much. I did resolve it so: first_time=1 exec < $DATA_HOME/$AMEXFILE while read -r line do head -1 $DATA_HOME/$AMEXFILE | awk '{print substr($0,1,112) " " substr($0,115,3 28)}' > formatted_amex1205.txt awk '{print substr($0,1,112) " " substr($0,115,328)}' >> formatte d_amex1205.txt done But as you can see I have a redundant processing for the first line with the head command. Some how my loop process all records except the first record. Any idea. lakb200

0

Wufgelazabarak! I will explain the problems with your code in a subsequent reply, but first, just for the sake of experiment, please do me a favor ... Copy/paste my first solution exactly as it is, except to change input and outfile filenames.  Change nothing else - just the file names - and see if the output is what you want.  Then copy/paste my second solution exactly as is, change only the file names, and see if that output is what you want. Let us know the results.  Thanks. In a short while, I will explain your code (whose behaviour is consistent with RedHat linux but not SuSE linux).

0

Your while statement reads its first line (from the defaulted stdin that was declared by the exec statement) into the $line variable (which your code does not utilize).  That first while loop iteration burns (eats) the first line of your input (and does nothing with it). Not to worry, because your head -1 infile | awk goes and gets that first line again (rather than utilize the copy sitting in $line), and sends it to the output file. This is immediately followed by another awk statement that does not specify an input file, which means that it, also, will read the standard input that was declared by your exec statement.  On HP-UX and SuSE linux, it appears that awk continues to read the remainder of the stdin that has already been opened by the while loop (and the first line having already been eaten). Therefore (and stranger things have been known to happen), this code actually produces the desired output on HP-UX and SuSE linux. However, on RedHat linux (and probably others), that second awk appears to open its own second copy of the (defaulted) stdin, so it processes the entire input file (including the first line).  So, your duplication is due to the head -1 infile | awk outputting record 1, then the second awk outputs the entire file again, including another record 1. Note that all of this is happening on the first iteration of the while loop. That second awk statement consumes the entire input file. Decide and visualize how you want to process a file.  Commands that process a file will retrieve all of the lines on their own, so they don't need to be inside a loop.  Or instead, you can use a "while read" loop to read and process one line at a time.  In that case, you will read each input record into a variable (read line) or maybe read into several variables (read word1 word2 word3), then you can test or manipulate those variables and maybe output them with echo or print. Inside a while loop that is reading file1, you typically will not use commands that also read file1.

0

Thanks so much to identify the mystery. Jim the code that you gave does work. I appreciate all the insight you gave us. lakb200

0

karabazalegfuW! what am i missing?: "abc dfdsfu dfsfymsa fafda OUTPUT abc dfdsfu dfsf sa fafda ... only change is ... subsituting dfsf sa for dfsfymsa" sed 's/dfsfymsa/dfsf sa/' Origfile>NewFile

0

That was just a sample line. The code needs to change columns 113-114 to spaces in each line, regardless of content.

0