I'm new with Pandas. I have the following dataframe.
Group type G1 a1 G1 a2 G1 a3 G2 a2 G2 a1 G3 a1 G4 a1 G5 a4 G5 a1
And I would like to obtain for each couple of groups how many "types" they have in common. Something like this:
Group type count G1 a1 G1 a2 G1 a3 G2 a2 G2 a1 G3 a1 G4 a1 G5 a4 G5 a1 count: (G1, G2, 2) (Elements in common: a1,a2) count: (G1, G3, 1) (Elements in common: a1) count: (G1, G4, 1) (Elements in common: a1) ...
Do you have any idea how could I implement this? Is there any function from the pandas library that could guide me into the right direction.
I think you need numpy.intersect1d
:
import itertools
#get all combinations of Group values
c = list(itertools.combinations(list(set(df['Group'])), 2))
df = df.set_index('Group')
#create list of tuples of intersections and lengths
L = []
for a, b in c:
d = np.intersect1d(df.loc[a], df.loc[b]).tolist()
L.append((a,b, len(d), d))
#new DataFrame
df = pd.DataFrame(L, columns=['a','b','lens','common'])
print (df)
a b lens common
0 G2 G4 1 [a1]
1 G2 G1 2 [a1, a2]
2 G2 G3 1 [a1]
3 G2 G5 1 [a1]
4 G4 G1 1 [a1]
5 G4 G3 1 [a1]
6 G4 G5 1 [a1]
7 G1 G3 1 [a1]
8 G1 G5 1 [a1]
9 G3 G5 1 [a1]
Given a dataframe:
import pandas as pd
df = pd.DataFrame([['G1', 'G1', 'G2', 'G2'], ['a1', 'a2', 'a1', 'a3']]).T
df.columns = ['group', 'type']
Then there are two options:
df.groupby('type').count()
or if you want to know them explicitly:
df.groupby(['type', 'group']).count()
Thus you can do, eg:
df1.loc['a1']
with output:
group
G1
G2
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.