Okay... pythons build-in methods are quite fast. so is hex().
with about 64 kb memory i can write it with a 16 bit dictionary
where the dictionary generation itself is not yet optimized:
def genBitList(exp):
next = lambda now: [x+'0' for x in now]+[x+'1' for x in now]
result = [""]
for x in range(exp):
result = next(result)
return result
_digits = genBitList(16)
def bitstring2(number):
"""lsb------>msb"""
rlist = list()
if number >= 0:
while number:
rlist.append( _digits[number & 0xFFFF] )
number >>= 16
return ''.join(rlist).rstrip('0')
else:
while number != -1:
rlist.append( _digits[number & 0xFFFF] )
number >>= 16
return ''.join(rlist).rstrip('1')
this is quite fast and in lots of cases faster than the
hex()-version. however your method (with my inverses formatting) is
very fast, too and without so much memory expense:
_nibbles = {"0":"0000", "1":"0001", "2":"0010", "3":"0011",
"4":"0100", "5":"0101", "6":"0110", "7":"0111",
"8":"1000", "9":"1001", "a":"1010", "b":"1011",
"c":"1100", "d":"1101", "e":"1110", "f":"1111",
"l":"", "-":"", "x":""}
from string import maketrans, translate
_invert = maketrans('01', '10')
def bitstring3( number ):
if number > 0:
return ''.join( [_nibbles[d] for d in
hex( number ).lower()] ).lstrip( '0' )
else:
return translate(''.join( [_nibbles[d] for d in
hex( ~number ).lower()] ).lstrip( '0' ), _invert)
Benchmark:
for 0xa:
bitstring2: 0.61802315712
bitstring3: 0.91001200676
for 0xaaaa:
bitstring2: 0.561501026154
bitstring3: 1.11787199974
for 0xaaaaaaaaaaaa:
bitstring2: 1.2295820713
bitstring3: 1.61559510231
for 0xaaaaaaaaaaaaaaaaaaaaaaaa:
bitstring2: 1.90036797523
bitstring3: 2.2683339119
for -0xaaaaaaaaaaaaaaaaaaaaaaaa:
bitstring2: 2.81339716911
bitstring3: 2.74266886711
mfg
- eth
--
http://mail.python.org/mailman/listinfo/python-list
