Value of Floating Point Fields

frankvijt · New User Joined: 18 Jan 2008 Posts: 2 Location: belgium

I am working on a conversion project, which involves reading out a whole lot of good old VSAM files , which contain valuable data.

The original programming language, used to store these data, is PL/I. Some of the fields have been defined as FLOAT(6).

As such that is no problem, as I just dump the data via PUT SKIP EDIT statements onto flat files, that then get converted to txt files for our PC wiz kids to do their evil stuff with.

I do however want to make sure that the wiz kids get the right data. I can display the data in the files in hex, which is fine for reading and checking binary or dec fields. For Float however it isn't as straight forward.

For example, I have a field which is defined FLOAT(6) and has the hexadecimal value :
4100
1000

I believe this to be the representation of 1.00000

Some convertor, that I found on the net, however tries to tell me that it represents the value 9.00000, which got me confused.

Does any one have any experience with this and how did you handle it?
Thanks!

Robert Sample · Posted: Mon Dec 08, 2008 9:58 pm

It has been a while, but floating point numbers are expressed in terms of a sign bit, biased exponent for 7 bits (hex 40 is the normal bias, so 41 means the number is times 10), and the remaining bits represent the value. I've confirmed via SAS that 41100000 is 1.E0 floating point.

frankvijt · New User Joined: 18 Jan 2008 Posts: 2 Location: belgium

Thanks Robert for the quick answer and the additional extra information!
That solves my query perfectly.

Robert Sample · Posted: Mon Dec 08, 2008 11:34 pm

Glad to hear it.

avina · New User Joined: 04 Dec 2007 Posts: 16 Location: Kolkata

Thanks a lot for your reply. It helps a lot.