Generate combinations from list within column

Question

We have a DataFrame with 2 columns as follows:

|Type   |list_dates  |  
|:----:|:-----------:|  
|1    |['a','b','c']|  
|2    |['d','e','f','g']|

We need to generate a combination of all list elements while duplicating the Type, as follows:

|Type   |list_dates  |  
|:----:|:-----------:|  
|1    |['a','b']|  
|1    |['a','c']|  
|1    |['b','c']|  
|2    |['d','e']|  
|2    |['e','f']|  
.....

In order to generate the combinations we are using the following code:

import itertools
list(itertools.combinations(df.list_dates,2) )

Any suggestions?

jezrael · Accepted Answer · 2017-11-20 10:29:17Z

1

I think pure python solution working best. So first create tuples by dict and then create list of tuples by combinations. Last create DataFrame by constructor:

import itertools

L = []
for x, y in zip(df['Type'], df['list_dates']):
    a = list(itertools.combinations(y,2))
    for i in a:
        L.append((x, list(i)))

Or nested list comprehension solution:

L = [(x, list(i)) for x, y in zip(df['Type'], df['list_dates']) 
                  for i in list(itertools.combinations(y,2))]

df = pd.DataFrame(L, columns=['Type','list_dates'])
print (df)

   Type list_dates
0     1     [a, b]
1     1     [a, c]
2     1     [b, c]
3     2     [d, e]
4     2     [d, f]
5     2     [d, g]
6     2     [e, f]
7     2     [e, g]
8     2     [f, g]

Thanks, piRSquared for nice suggestion - if DataFrame have only 2 columns:

import itertools

L = []
for x, y in df.values:
    a = list(itertools.combinations(y,2))
    for i in a:
        L.append((x, list(i)))

L = [(x, list(i)) for x, y in df.values for i in list(itertools.combinations(y,2))]

If more columns first filter:

L = [(x, list(i)) for x, y in df[['Type','list_dates']].values 
                  for i in list(itertools.combinations(y,2))]

edited Nov 20, 2017 at 10:29

answered Nov 20, 2017 at 10:07

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

piRSquared Over a year ago

You don't need the zip pd.DataFrame([[x, list(y)] for x, z in df.values for y in itertools.combinations(z, 2)], columns=['Type','list_dates'])

skibee Over a year ago

Great! Thank you

Collectives™ on Stack Overflow

Generate combinations from list within column

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related