CODEWITHSUNDEEP
pythonpandas
PythonF
PythonF

Reputation: 496

Pandas isin with multiple columns

I want to select all rows in a dataframe which contain values defined in a list. I've got two approaches which both do not work as expected/wanted.

My dataframe looks something like this:

Timestamp DEVICE READING VALUE
1 | DEV1 | READ1 | randomvalue
2 | DEV1 | READ2 | randomvalue
3 | DEV2 | READ1 | randomvalue
4 | DEV2 | READ2 | randomvalue
5 | DEV3 | READ1 | randomvalue

and I've got the list (ls) like follows:

[[DEV1, READ1], [DEV1, READ2], [DEV2,READ1]]

In this scenario I want to remove line 4 and 5:

My first approach was:

df = df[(df['DEVICE']. isin([ls[i][0] for i in range(len(ls))])) &
        (df['READING'].isin([ls[k][1] for k in range(len(ls))]))]

The problem with this one is obviously, that it does not remove line 4, because DEV2 has the READING READ2, but it should remove it.

My second approach was:

df = df[(df[['DEVICE','READING']].isin({'DEVICE':  [ls[i][0] for i in range(len(ls))],
                                        'READING': [ls[i][1] for i in range(len(ls))] }))]

This one selects the correct rows but it does not remove the other rows. Instead it sets every other cell to NaN, including the VALUE ROW, which i do want to keep. And It does not accumulate both so row 4 looks like 4 |DEV2|NaN|NaN

What would be the easiest or best way, to solve this problem? Can you help me?

~Fabian

Upvotes: 8

Views: 13843

Answers (4)

Matthias Ossadnik
Matthias Ossadnik

Reputation: 911

You can use a multi-index to solve this problem.

values = [['DEV1', 'READ1'], ['DEV1', 'READ2'], ['DEV2', 'READ1']]
# DataFrame.loc requires tuples for multi-index lookups
index_values = [tuple(v) for v in values]

filtered = df.set_index(['DEVICE', 'READING']).loc[index_values].reset_index()
print(filtered)

  DEVICE READING  Timestamp        VALUE
0   DEV1   READ1          1  randomvalue
1   DEV1   READ2          2  randomvalue
2   DEV2   READ1          3  randomvalue

Upvotes: 5

James Fulton
James Fulton

Reputation: 332

This should do what you want

import pandas as pd

df = pd.DataFrame({'a':[1,1,0,0,1], 'b':[0,0,1,0,1]})

keepers = [[0,0],[1,1]]

df = df[df.apply(lambda row: [row['a'], row['b']] in keepers, axis=1)]

Upvotes: 1

Vaishali
Vaishali

Reputation: 38415

You can convert the list to list of tuples. Convert the required columns in dataframe to tuples and use isin

l = [['DEV1', 'READ1'], ['DEV1', 'READ2'], ['DEV2','READ1']]
l = [tuple(i) for i in l]
df[df[['DEVICE', 'READING']].apply(tuple, axis = 1).isin(l)]

You get

    Timestamp   DEVICE  READING VALUE
0   1   DEV1    READ1   randomvalue
1   2   DEV1    READ2   randomvalue
2   3   DEV2    READ1   randomvalue

Upvotes: 15

kpie
kpie

Reputation: 11100

Any reason you don't do it like this?

df.drop([4,5],axis=0,inplace=True)

Upvotes: -1

Related Questions