Python：嵌套的iteritems循环仅运行一次

Question

I'm trying to loop over a dictionary, and append dictionary inside it. But when I use loop using iteritems(), it just iterating once and exit to the outer loop. I am quite new in python but I think this code should be working.

Why this happen?

from __future__ import division
import numpy
import pprint

training = {"outlook":{"sunny":{"yes":2,"no":3},"overcast":{"yes":4,"no":0},"rainy":{"yes":3,"no":2}},\
            "temperature":{"yes":[83,70,68,64,69,75,75,72,81],"no":[85,80,65,72,71]},\
            "humidity":{"yes":[86,96,80,65,70,80,70,90,75],"no":[85,90,70,95,91]},\
            "windy":{"false":{"yes":6,"no":2},"true":{"yes":3,"no":3}},\
            "play":{"yes":9,"no":5}}
processed = {}
total_yes = training["play"]["yes"]
total_no = training["play"]["no"]
total_all = training["play"]["no"]+training["play"]["yes"]

def main():
    for k,v in training.iteritems():
        if(k != "play"):
            for k2,v2 in v.iteritems():
                if((k2 == "yes") & isinstance(v2,list)):
                    processed[k] = {"yes":{"mean":numpy.mean(v2),"std":numpy.std(v2,ddof=1)}}
                elif((k2 == "no") & isinstance(v2,list)):
                    processed[k].update({"no":{"mean":numpy.mean(v2),"std":numpy.std(v2,ddof=1)}})
                else:
                    processed[k] = {}

                    # the problems start here
                    for i,j in v2.iteritems():
                        if(i == "yes"):
                            p_yes = j/total_yes
                            processed[k].update({k2:{"yes":p_yes,"no":(1-p_yes)}})
                        #when I print processed[k], it contains current value only (not including previous value)
                #suddenly exit after iterating once

        else:
            processed[k] = {"yes":total_yes/total_all,"no":total_no/total_all}

    pp = pprint.PrettyPrinter(indent=3)
    pp.pprint(processed)

if __name__ == '__main__':
    main()

The output is :

...
'outlook': {  'sunny': {  'no': 0.7777777777777778,
                          'yes': 0.2222222222222222}},
...

But expected output is :

...
'outlook': {  'sunny': {  'no': 0.7777777777777778,
                          'yes': 0.2222222222222222}},
              'overcast': { 'no':xxxxx,
                            'yes':xxx
                          }
              ...
...

Answer 1

You're mistake was that you always erase the 'outlook' and in the end you see the last entry that was inserted into 'outlook' dictionary. Here is a fix of your code:

from __future__ import division
import numpy
import pprint

training = {"outlook":{"sunny":{"yes":2,"no":3},"overcast":{"yes":4,"no":0},"rainy":{"yes":3,"no":2}},\
        "temperature":{"yes":[83,70,68,64,69,75,75,72,81],"no":[85,80,65,72,71]},\
        "humidity":{"yes":[86,96,80,65,70,80,70,90,75],"no":[85,90,70,95,91]},\
        "windy":{"false":{"yes":6,"no":2},"true":{"yes":3,"no":3}},\
        "play":{"yes":9,"no":5}}
processed = {}
total_yes = training["play"]["yes"]
total_no = training["play"]["no"]
total_all = training["play"]["no"]+training["play"]["yes"]


def main():
    for k,v in training.iteritems():
        if (k=='outlook'):
            pass
        if(k != "play"):
            for k2,v2 in v.iteritems():
                if((k2 == "yes") & isinstance(v2,list)):
                    processed[k] = {"yes":{"mean":numpy.mean(v2),"std":numpy.std(v2,ddof=1)}}
                elif((k2 == "no") & isinstance(v2,list)):
                    processed[k].update({"no":{"mean":numpy.mean(v2),"std":numpy.std(v2,ddof=1)}})
                else:
                    if k not in processed:
                        processed[k] = {}

                    for i,j in v2.iteritems():
                        if(i == "yes"):
                            p_yes = j/total_yes
                            processed[k].update({k2:{"yes":p_yes,"no":(1-p_yes)}})                       


        else:
            processed[k] = {"yes":total_yes/total_all,"no":total_no/total_all}

    pp = pprint.PrettyPrinter(indent=3)
    pp.pprint(processed)

if __name__ == '__main__':
    main()

Answer 2

The problem lies here:

else:
    processed[k] = {}

The previous value of k is overwritten as your for loop moves to the next element.

Define this here and your problem should be solved:

for k,v in training.iteritems():
    if(k != "play"):
        processed[k] = {}