I have been trying to convert a set of 2 16 bit integers into a single 32 bit floating point number. This is for my work on MODBUS communication protocol. My query is the same as mentioned here
However, when I follow the pack/unpack approach provided there, I do not get similar results.
Here is my test program:
from struct import *
tup=(16256, 0)
print hex(tup[0])
print hex(tup[1])
mypack = pack('>HH',tup[0],tup[1])
print mypack
f = unpack('f', mypack)[0]
print f
And the Output is as below:
0x3f80
0x0
?
4.60060298822e-41
I expect the final output to be actually just '1.0' as the floating point interpretation of '0x3F800000' is just that!
I am running this code on a Beagle Bone Black board, powered by Debian OS.
Where am I going wrong?
You need to specify the endianness in your unpack call.
from struct import *
tup=(16256, 0)
print hex(tup[0])
print hex(tup[1])
mypack = pack('>HH',tup[0],tup[1])
print `mypack`
f = unpack('>f', mypack)
print f
output
0x3f80
0x0
'?\x80\x00\x00'
(1.0,)
But please bear in mind the warnings / caveats mentioned in the question you linked to, especially those in regard to NaN.
Also note that it's generally considered bad practice to do from modulename import * because it dumps the names from modulename into your scripts namespace. This is messy, and it can lead to bugs. So it's better to do
import struct
#...
mypack = struct.pack('>HH',tup[0],tup[1])
f = struct.unpack('>f', mypack)
Sure, it's a little bit more to type, but it makes your code a lot easier to read and maintain.