CODEWITHSUNDEEP
python-3.xpandasdataframe
Nelly Louis
Nelly Louis

Reputation: 167

How can I drop rows in data frames which contains empty lists?

I have created a data frame with 3 columns, the third one contains lists, I want to drop rows that contains an empty list in that cell.

I have tried with

df[df.numbers == []] and df[df.numbers == null]
but nothing works. 

name    country    numbers

Lewis   Spain      [1,4,6]
Nora    UK         []
Andrew  UK         [3,5]

The result will be a data frame without Nora's row

Upvotes: 9

Views: 4599

Answers (6)

Chris Nielsen
Chris Nielsen

Reputation: 849

Let's say your data is set up like this:

import pandas as pd
df = pd.DataFrame([{'name': "Lewis", 'country': "Spain", "numbers": [1,4,6]}, 
                   {'name': "Nora", 'country': "UK", "numbers": []},
                   {'name': "Andrew", 'country': "UK", "numbers": [3,5]}])

You could iterate over the dataframe and add only the rows that don't have an empty numbers array to a new dataframe called "newDF". For example:

newDFArray = []
for index, row in df.iterrows():
    emptyArrayCheck = row["numbers"]
    if len(emptyArrayCheck) > 0:
        newDFArray.append(row)
newDF = pd.DataFrame(newDFArray)

newDF

This will yield:

    country name    numbers
0   Spain   Lewis   [1, 4, 6]
2   UK      Andrew     [3, 5]

Upvotes: 3

tawab_shakeel
tawab_shakeel

Reputation: 3739

just check len > 0

df[df['numbers'].str.len()>0]

Upvotes: 7

BENY
BENY

Reputation: 323226

Umm check bool

df[df.numbers.astype(bool)]

Upvotes: 17

IWHKYB
IWHKYB

Reputation: 491

One way to do it is to create a new column containing the length of df.numbers by:

df['len'] = df.apply(lambda row: len(row.numbers), axis=1)

and then filter by that column by doing:

df[df.len > 0]

Upvotes: 4

anky
anky

Reputation: 75080

Use series.str.len() to check the length of elements in the list and then filter out where it equals 0:

df[~df.numbers.str.len().eq(0)]

     name country    numbers
0   Lewis   Spain  [1, 4, 6]
2  Andrew      UK     [3, 5]

Upvotes: 13

sjw
sjw

Reputation: 6543

Using the idea that the result of any list multiplied by 0 gives an empty list, one way to do this is:

In [29]: df[df.numbers != df.numbers * 0]

Out[29]: 
     name    numbers country
0   Lewis  [1, 4, 6]   Spain
2  Andrew     [3, 5]      UK

Upvotes: 5

Related Questions