Parse/unpivot data into the column in pandas dataframe

Question

I have a pandas DataFrame with the following structure:

|A|B|C      |
-------------
|1|2|"1,2,4"|
|3|4|"5,6"  |

What is the most pythonic way to get a table like below?

|A|B|C|
-------
|1|2|1|
|1|2|2|
|1|2|4|
|3|4|5|
|3|4|6|

The initial table contains ~10K instances with 4-18 elements in C .

For the moment I have nothing brighter than:

import pandas as pd
df = pd.DataFrame([[1, 2, "1,2,4"], [3, 4, "5,6"]], columns=['A', 'B', 'C'])
tmp = df['C'].str.split(',').apply(pd.Series, 1).stack()
tmp.index = tmp.index.droplevel(-1)
tmp.name = 'C'
del df['C']
df = df.join(tmp)

Answer 1

Not much better. I did it two ways. The first is similar to yours with some cosmetic differences and feels like the better of the two. I still added the second because it's a bit different in that it concats the series returned by split.

>>> import pandas as pd
>>> df = pd.DataFrame([[1, 2, "1,2,4"], [3, 4, "5,6"]], columns=['A', 'B', 'C'])
>>> x = df.drop('C', axis=1).join(df.C.str.split(',', expand=True).stack().to_frame('C').reset_index(1, drop=True))
>>> print x
   A  B  C
0  1  2  1
0  1  2  2
0  1  2  4
1  3  4  5
1  3  4  6

>>> y = df.drop('C', axis=1).join(pd.concat(df.C.str.split(',', expand=True).to_dict('series').values()).dropna().to_frame('C'))
>>> print y
   A  B  C
0  1  2  1
0  1  2  2
0  1  2  4
1  3  4  5
1  3  4  6

Answer 2

you can do this by for loop:

with open("your_file") as f:
    for i,line in enumerate(f):
        if i==1:
            print line
        else:
            line = line.split("|")
            elements = line[-1].strip("\" ").split(",")
            for x in elements:
                print "|{}|{}|".format("|".join(line[:2]), x)