I am attempting to pack the contents of a python bytearray into a 4byte signed integer using struct.pack. Unfortunately, pack wants a string, so after some googling I figured I needed to decode my bytearray to a string. I figured ascii meant since because an ascii character is a byte long. Unfortunately, ascii did not want to support my values > 127, so I thought I would use replace...
but when I do this decode returns an object of type unicode and now each of my bytes is a 4 character string...
This just seems a little ridiculous, im missing something obvious (ps I have been using python for about two weeks)
here is what I am trying to do...
val = long(-5)
s = bytearray(pack("<i", val))
s.pop() # pop off msb
# write it out the way we want to then read it in the way the code does
fout = open("test.bat", "wb")
fout.write(s)
fout.close()
fin = open("test.bat", "rb")
inBytes = bytearray(fin.read(3))
# extend sign bit
if (inBytes[2] & 0x80):
inBytes.append(0xff)
else:
inBytes.append(0x00)
nb = inBytes.decode('ascii', 'replace')
# ERROR:root:after decode, len: 4 type: <type 'unicode'>
logging.error("after decode, len: {0} type: {1}".format(len(nb), type(nb)))
# struct.error: unpack requires a string argument of length 4
inInt32 = unpack('<i', inBytes.decode('ascii', 'replace'))[0]
fin.close()
When you read from a file in binary mode, you get an object that can be used immediately with
struct.unpack
.Creating the input data:
Python 2.x .. it's a
str
object.Python 3.x ... it's a
bytes
object.All you need is to cast
inBytes
back tostr
: