Add stack argument to str.findall

[hayd]

Some of the string methods return a Series of lists (e.g. findall) which isn't that convenient for doing more analysis:

df = pd.DataFrame([['@a @b'], ['@a'], ['@c']], columns=['tweets'])
at_mentions = df.tweets.str.findall('@[a-zA-Z0-9_]+')

In [3]: at_mentions
Out[3]:
0    [@a, @b]
1        [@a]
2        [@c]
Name: tweets, dtype: object

I think it would be nice to be able to stack the results:

In [4]: at_mentions.apply(pd.Series).stack()
Out[4]:
tweets
1       0    @a
        1    @b
2       0    @a
3       0    @c
dtype: object

see this SO answer

Could just do it directly like that, or perhaps make more efficient. Thoughts?

[jreback]

I actually like this (as the default)

In [7]: at_mentions.apply(lambda x: Series(1,x)).fillna(0)
Out[7]: 
   @a  @b  @c
0   1   1   0
1   1   0   0
2   0   0   1

[cpcloud]

clever....that's the sweet dummy variable hack

[jreback]

realted is #4685, pushing to 0.14

[mroeschke]

I think explode solves this use case. Happy to reopen if this doesn't solve the original use case.

In [7]: at_mentions.explode()
Out[7]:
0    @a
0    @b
1    @a
2    @c
Name: tweets, dtype: object