IEEE 754 conversion in Python and reading individual bytes

Question

Hi,
1. If i wanted to just read a particular number of bytes, for example, byte 4-8 in a 15 byte data stream, how would i do that in python please?

2. Does anybody know how to convert a 32 bit hex data into IEEE 754 floating point in python please!

Answer 1

> 1. If i wanted to just read a particular number of
> bytes, for example, byte 4-8 in a 15 byte data
> stream, how would i do that in python please?

Once you've read your 15 bytes into a string, you can use slices to extract the sub-string that you desire.

For example:
a = 'abcdefghijklmnp'
b = a[4:9]

will extract the bytes with subscripts 4 through 8 inclusive.

>
> 2. Does anybody know how to convert a 32 bit hex data
> into IEEE 754 floating point in python please!

IEEE 754 has more than one 32 bit representation for data. However, I'll assume you mean something like a single precision float. If that's the case check out.

http://en.wikipedia.org/wiki/Binary32

to make sure. Probably the simplest way is to use the struct module to unpack it as a float. If you have your 32 bits as a binary string in variable h, you would just do:

struct.unpack('f', h)

and it would return the data as a float object.

However,you mention that it is "hex" data. So if it's a formatted string represented in hex, you'll need to get the data into a packed binary representation first. One way would be to use the built in int function int(h, 16) for example, and then unpack that.

Answer 2

Remember that unlike C/C++, where a "ASCII string" is a null-terminated thing of variable length and a "binary string" is a byte array of fixed length, in Python both are just "strings" which can be queried for length or segments as Digiguy says above.

So if you receive 15-bytes from a serial port or UDP/ZB socket, it is a "string" regardless of being text or binary. I mention this because some Python beginners get this mind-set that binary "byte arrays" are somehow different than strings. But st = '\x00\x01\x02\x03' creates a 4-byte string just as does st = 'ABCD'.

(Side-Note: yes, some C/C++/API into Python pass in a raw non-Python "byte buffers", but those are not common outside of things like raw Windows MFC/API linkages).

Answer 3

Thank you to both of you! That was very helpful!

Answer 4

hi, i was wondering if you can help me again please,
i'm reading raw hex data and i want to be able to extract particular bytes and display its decimal equivalent.
For example, if the raw data is: aa1084e89d41000000001e ...
i want to extract out the following 4 bytes: 84e89d41
and then i want to read it as 419de884 (so LSB is 8 and MSB is 1 i presume, and each byte/pair is reversed). How can i do this please?

Sorry for such basic questions .. i'm no good with programming.

Answer 5

hi, i was wondering if you can help me again please,
i'm reading raw hex data and i want to be able to extract particular bytes and display its decimal equivalent.
For example, if the raw data is: aa1084e89d41000000001e ...
i want to extract out the following 4 bytes: 84e89d41
and then i want to read it as 419de884 (so LSB is 8 and MSB is 1 i presume, and each byte/pair is reversed). How can i do this please?

Sorry for such basic questions .. i'm no good with programming.

Answer 6

I first assume that your representation is actually raw octets and *not* the letter "A" followed by the letter "A" followed by "1", etc.

Look at the "struct" module.

Here is an example:

data = "\xAA\x10\x84\xE8\x9D\x41\x00\x00\x00\x00\x1E"
print len(data)

# One way to get the value from unpack
values = struct.unpack('>L', data[2:6])
value = values[0]
print "%08x -- %d" % (value, value)

# Another way to get the value
value, = struct.unpack('<L', data[2:6])
print "%08x -- %d" % (value, value)