Converts hexadecimal positive binary values to either integer
(fixed-point) or real (floating-point) binary values
| Numeric |
|
| Width range: |
1- 16
|
| Default width: |
8
|
| CMS specifics: |
IBM floating-point format
|
-
w
-
specifies the field width of the input value
and determines whether the input represents an integer or real binary value.
Each hexadecimal digit that is read by the HEX informat
is interpreted using the EBCDIC character-encoding system, with one digit
per byte. For example, the hexadecimal value 'F1'x (EBCDIC 1) results in the
number 1.0 using the HEX1. informat.
When you specify a width value of 1 through 15, the
input hexadecimal number represents an integer binary value. When you specify
16 for the field width, the input hexadecimal number represents a real binary
number.
For example, suppose a floating-point number has been
stored as the following hexadecimal character value:
433E800000000000
('F4F3F3C5F8F0F0F0F0F0F0F0F0F0F0F0'x)
If you read the value using HEX16. as the informat,
the number is converted to its correct floating-point value of 1000. However,
if you specify HEX15. as the informat, SAS expects the input to represent
an integer binary value. Since the number was originally stored in IBM floating-point
format, the second result is incorrect.
If a number has been stored as a character representation
of an integer binary value, it is correct to use a w value of
less than 16. For example, if 256 has been stored as the following hexadecimal
character value, then using HEX6. gives the correct result:
000100 ('F0F0F0F1F0F0'x)
However, using HEX16. produces an incorrect result of
3.38E-80 in the preceding example.
You can use the HEX informat to read negative floating-point
numbers; however, it cannot be used to read negative integer binary numbers.
For example, if you write -10 using the HEX16. format, it is stored in a
character representation of a floating-point format, as follows:
C1A0000000000000
('C3F1C1F0F0F0F0F0F0F0F0F0F0F0F0F0'x)
If this character representation of a floating-point
number is then read with the HEX16. informat, the value is correctly interpreted
as the number -10.
However, if you write -10 as an integer binary number
using the HEX2. format, it is written in EBCDIC as the hexadecimal string
F6, which is the twos complement notation for -10. If you then read this value
using HEX2. as the informat, it is read as a positive number, giving a result
of
246.
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.
