I want to check all rows in dataframe,

• if this row's column1 == previous row's column2

  and

• if this row's column2 == previous row's column1

then drop this row.

For example:

 Column1    Column2 
    A          B
    B          A
    A          C
    A          D
    D          A
    B          C

The result I want would be:

 Column1    Column2 
    A          B
    A          C
    A          D
    B          C

Please help, thank you!

CodePudding user response：

You can use

m1 = df['Column1'] == df['Column2'].shift()
m2 = df['Column2'] == df['Column1'].shift()

out = df[~(m1 & m2)]

print(out)

  Column1 Column2
0       A       B
2       A       C
3       A       D
5       B       C

CodePudding user response：

You could do a shift on those columns then use conditional statement to check if the column values are equal.

import pandas as pd


data = [
    ['A',          'B'],
    ['B',          'A'],
    ['A',          'C'],
    ['A',          'D'],
    ['D',          'A'],
    ['B',          'C']]

cols = ['Column1', 'Column2']

df = pd.DataFrame(data, columns=cols)

df['Pre_Column2'] = df['Column2'].shift(1)
df['Pre_Column1'] = df['Column1'].shift(1)

df = df[~((df['Column1'] == df['Pre_Column2']) & (df['Column2'] == df['Pre_Column1']))]
df = df[cols]

Output:

print(df)
  Column1 Column2
0       A       B
2       A       C
3       A       D
5       B       C