Python将旁串转换为字典

Question

The format of string is like "a:1 b:2 c:xd:2.13e-5" , is there some way to convert it to python dict quickly and simply?

-------------- edit line --------------

According the great answers, I tried several methods (in ipython):

In [6]: import re

In [7]: %paste
def f1(line):
    item_dict = {}
    for item in line.split():
        kv = item.split(':')
        item_dict[kv[0]] = kv[1]

def f2(line):
    item_dict = {}
    item_pat = re.compile(r'(\w+):(.+)')
    for item in line.split():
        m_res = item_pat.match(item)
        item_dict[m_res.group(1)] = m_res.group(2)

def f3(line):
    dict(item.split(':') for item in line.split())
## -- End pasted text --

In [8]: line = 'a:1   b:3243 dsfds:4323llsjdf         \t fdsf:3232l'
In [9]: %timeit f1(line)
100000 loops, best of 3: 3.99 us per loop

In [10]: %timeit f2(line)
100000 loops, best of 3: 8.83 us per loop

In [11]: %timeit f3(line)
100000 loops, best of 3: 5.19 us per loop

The first method f1() seems faster, but in my application, it still use much time(about 30% of all) because it's invoked millions of times.

Are there any more effective ways? Or cython ?

Answer 1

>>> s = "a:1 b:2 c:x d:2.13e-5"
>>> dict( p.split(':') for p in s.split(' ') )
{'a': '1', 'c': 'x', 'b': '2', 'd': '2.13e-5'}

This split s the string, first by spaces, to get the key-value pairs ( p ). Then it splits each pair by ':' to yield each key/value to be added to the dictionary.

Note that no conversion has taken place. All keys/values are still strings. To do any better than this, you're going to need a somewhat smart function that will convert any input string into your expected types.

Answer 2

import ast

def guess(s):
    try:
        return ast.literal_eval(s)
    except ValueError:
        return s    

s = "a:1 b:2 c:x d:2.13e-5"
print dict(map(guess, x.split(':')) for x in s.split())

---

{'a': 1, 'c': 'x', 'b': 2, 'd': 2.13e-05}