re-combining a list/dictionary in python

following o开发者_Go百科n from this question

i have the following lists in python which i want to recombine into a dictionary/list:

from

fromfruits = { "names" : ['banana','grapefruit','apple'] , "colors" : ['yellow','pink','green'], ... }

to

tofruits = [
{'name':'banana','color':'yellow',...}, 
{'name':'grapefruit','color':'pink',...},
{'name':'apple','color':'green',...}
]

what's the best way to do it so that i can have n properties listed in fromfruits?

please help! :)

---

Since the request is now to be more general:

>>> from itertools import izip
>>> ff = {'colors': ['yellow', 'pink', 'green'], 'names': ['banana', 'grapefruit', 'apple'], 'blah': ['a','b','c']}

>>> [dict(izip(ff.iterkeys(), v)) for v in izip(*ff.itervalues())]
[{'blah': 'a', 'colors': 'yellow', 'names': 'banana'},
 {'blah': 'b', 'colors': 'pink', 'names': 'grapefruit'},
 {'blah': 'c', 'colors': 'green', 'names': 'apple'}]

since the order of keys and values are the same (assuming no intervening modifications to the dictionary).

---

It'd be pretty hard to go from keys like 'names' to 'name', teaching the program how to do proper english singlularization ... so i renamed the keys in the input:

ff = dict(name=['banana','grapefruit','apple'], color=['yellow','pink','green'], 
          yummy=[True,False,True])

You can solve this problem with zip again:

# make fruits [('yellow', True, 'banana'), ('pink', False, 'grapefruit'), ... ]
fruits = zip(*ff.itervalues())

# then add the names to each fruit
tofruits = [dict(zip(ff.iterkeys(),fruit)) for fruit in fruits]
# gives: [{'color': 'yellow', 'yummy': True, 'name': 'banana'}, ... ]

---

[dict((x, fromfruits[x][n]) for x in fromfruits.keys())
  for n in range(len(next(fromfruits.itervalues())))]

Optimize as desired.

---

>>> fromfruits
{'colors': ['yellow', 'pink', 'green'], 'names': ['banana', 'grapefruit', 'apple']}
>>> [{'name': name, 'color': color} for name in fromfruits['names'] for color in fromfruits['colors']]
[{'color': 'yellow', 'name': 'banana'}, {'color': 'pink', 'name': 'banana'}, {'color': 'green', 'name': 'banana'}, {'color': 'yellow', 'name': 'grapefruit'}, {'color': 'pink', 'name': 'grapefruit'}, {'color': 'green', 'name': 'grapefruit'}, {'color': 'yellow', 'name': 'apple'}, {'color': 'pink', 'name': 'apple'}, {'color': 'green', 'name': 'apple'}]

And now in some more detail (re-formatted for clarity):

>>>[{'name': name, 'color': color} 
        for name in fromfruits['names'] 
        for color in fromfruits['colors']]

This is a "list comprehension" with a double for which goes over all combinations of names and colors. You can add a third loop if you want to mix in other attributes, like "shape" or whatever.

---

not very pythonic, but ...

keys = fromfruits.keys()
nvals = len(fromfruits[keys[0]])
tofruits = [ ]
for i in range(nvals):
    tofruits.append ({ })
    for k in keys:
        tofruits[-1][k] = fromfruits[k][i]

---

This problem is a bit more convoluted than the average list comprehension because of the nested dictionaries.

However if you can create the right iterable then a list comprehension is the way to go.

Try:

tofruits = [ {'name':n, 'color':c} for n,c in zip(fromfruits['names'], 
                                              fromfruits['colors']) ]

Here the zip function is used to produce tuples which match the correct name and color. This is a good way to go here because both are stored in basic lists within the fromfruits dict.

These tuples (from zip) are then unpacked into n and c, which are then used as in a typical list comprehension.