populating a nesting dictionary

Question

having trouble populating a nested dictionary and retaining previously populated keys. see this example:

fulldict={}
keys=['key1', 'key2', 'key3']

for key in keys:
    for i in xrange(3):
        x1 = np.random.randn(10)
        y1 = np.random.randn(10)
        fulldict[key] = {i:pd.DataFrame({'x1':x1, 'y1': y1})}

my intent is that fulldict['key1'] should contain 3 dictionaries with keys 0,1,2. but only the last key (2) is stored.

any suggestions appreciated

Answer 1

You are reassigning the fulldict[key] each time, so initialize fulldict[key] = {} and use i as a key:

for key in keys:
    fulldict[key] = {}
    for i in xrange(3):
        x1 = np.random.randn(10)
        y1 = np.random.randn(10)
        fulldict[key][i] = pd.DataFrame({'x1':x1, 'y1': y1})

Answer 2

Or you could resort to a dict comprehension for the inner dict (saves a few lookups):

import numpy as np

fulldict={}
keys=['key1', 'key2', 'key3']

draw = np.random.randn
for key in keys:
    fulldict[key] = {i: pd.DataFrame({'x1': draw(10), 'y1': draw(10)})
                     for i in xrange(3)}