Create pandas dataframe with unequal list of values

Question

Is there a built-in function apart from DataFrame.from_dict able to create a dataframe from a dictionary with unequal list of values?

What I am trying to do create a dataframe from the following dictionary:

d = {'g': {1, 2, 3}, 'h': {4, 5}}

so the output should look like:

   value  class
0    1        g
1    2        g
2    3        g
3    4        h
4    5        h

Is having the set instead of a list a problem?

score 6 · Answer 1 · answered May 12 '19 at 15:42

6

Yet another way using melt

pd.DataFrame(d.values(), d.keys()).T.melt().dropna()

answered May 12 '19 at 15:42

rafaelc

52,436
15
51
78

1

interesting... didn't know about melt. Thansk :) – Kenneth Rivadeneira Guadamud May 12 '19 at 15:47
1

@KennethRivadeneiraGuadamud you can consider upvoting the answers that helped you. :) – anky May 12 '19 at 15:50
I'd need more details than that ;] – rafaelc May 12 '19 at 16:36
I think it needs list. pd.DataFrame(list(d.values()), list(d.keys())).T.melt().dropna() I like this solution better. :) – run-out May 12 '19 at 16:37
@run-out I think depending on the pandas version you're using, you need to add `list()` you're right ;} – rafaelc May 12 '19 at 16:40
1

Ah. I see. Good answer. – run-out May 12 '19 at 16:42

score 4 · Accepted Answer · answered May 12 '19 at 15:37

If you don't mind manually unwrapping, you can achieve this through

In [9]: pd.DataFrame([(x, k) for k, v in d.items() for x in v], columns=['value', 'class'])
Out[9]:
   value class
0      1     g
1      2     g
2      3     g
3      4     h
4      5     h

score 0 · Answer 3 · answered May 12 '19 at 16:52

0

I am using unnesting

unnesting(pd.Series(d).apply(list).to_frame('name'),['name']).reset_index()
  index  name
0     g     1
1     g     2
2     g     3
3     h     4
4     h     5

answered May 12 '19 at 16:52

BENY

296,997
19
147
204

Create pandas dataframe with unequal list of values

3 Answers3