IBM Mainframe Forum

Thanks for letting us know.

Perhaps you could post the AWK code, and some instructions on how to run it. Many on z/OS don't know that it is an option for processing things.

BillyBoyo wrote:Thanks for letting us know.

Perhaps you could post the AWK code, and some instructions on how to run it. Many on z/OS don't know that it is an option for processing things.

Sure. Here is a snippet:

Sample input:


The AWK script I've finally used is a bit more complicated because it doesn't generate the SQL statements itself, but only a list of variables that are used on a later JCL step to substitute their occurrences in a template, but the idea was to generate the SQL statement automatically calculating the length of the input tags. Here is a snippet that shows how to calculate the XML tag, which is what I wanted:


Note: The trim() function is not a built-in function in AWK, but it's easy to find an implementation in Internet.

Assuming the AWK script is in /tmp/gensql.awk and the sample input file is /tmp/input.txt, the way we would call the script from a Unix session is the following:


Where:

1) -v FS="," tells AWK to use the comma as the field separator
2) -v OFS="\n" tells AWK to use the newline character as the output field separator

This is the equivalent JCL step to call the AWK script.


I hope this idea helps anybody.
Kind regards.

I didn't find the username/password, but I did. Sorry for being misleading with my Id.
Regards.

Hi Billy,
Could you plz explain your solution. Specifically, build

The BUILD is the least interesting part, because all it is doing is the final calculation (plus the formatting).

It works by appending a sequence of hexadecimal digits at a particular position. After the SQZ with SHIFT=LEFT has removed all the trailing blanks, the number which is now at the position which was previously the start of the sequence tells you the number of bytes which were "squeezed" out. Then the is used as a source item for the calculation (length-of-field minus bytes-squeezed-out is equal to the length of the data).

Take this example, where x represents a blank. Position 10 has the sequence of numbers:



Now remove the blanks:



And position 10 contains 4, the number of blanks removed.

Hi Bily,
thanks for your explanation, I just want to understand the use of "Binary Zero". Why we need that.

Here was the original explanatory text: "and prefix the first quote by "something" (I chose a binary zero) to avoid confusion from consecutive quotes".

Here's the code:



If you look, the TRAIL=C'"' causes the 81,80 field to end in a quote after it is "squeezed". I need to make the next field quoted, so need to start with a C'"' (the value of A-QUOTE). Without something to intervene, there would be consecutive quotes, one from the TRAIL and one I inserted. I don't want that, so I can put anything, except a quote, before the quote I insert. The value I chose was a binary zero, but it could have been anything - except a quote.