CODEWITHSUNDEEP
pythonpython-3.xpandas
manosbar
manosbar

Reputation: 316

Filter pandas data frame based on previous and next values

My data frame has 3 columns [C1-C3]

C1  C2  C3
a   h   1
b   i   0
c   j   0
d   k   1
e   l   1
f   m   1
g   n   0

I am interested in filtering its rows based on criteria regarding the previous and next values Specifically, I would like to keep all rows, where the C3 has the pattern 0-1-1 (previous_value = 0 and next_value = 1). Thus, in the example that is used above, the data frame should be transformed into:

C1  C2  C3
d   k   1

Upvotes: 2

Views: 70

Answers (2)

sash_wash
sash_wash

Reputation: 146

With .shift() method, you can shift index by desired number of periods.

df[(df.C3.shift(1)==0) & (df.C3==1) & (df.C3.shift(-1)==1)]

    C1  C2  C3
3   d   k   1

Upvotes: 0

BENY
BENY

Reputation: 323226

Using shift 

m = df.C3.shift(1).eq(1) & df.C3.eq(1) & df.C3.shift(-1).eq(1)
0    False
1    False
2    False
3    False
4     True
5    False
6    False
Name: C3, dtype: bool
df = df[m]

Upvotes: 2

Related Questions