Inserting New Columns and respective values (on condition) in a Pandas DataFrame

Question

I have the following sample DataFrame and List (extracted from df).

import pandas as pd
color_list = ['green','blue','red','yellow','black']

df = pd.DataFrame({'object': ['car','plane','tree','house','phone'],
                   'colors': ['red, blue',
                              'red, yellow, black',
                              'black',
                              'green, blue',
                              'yellow, green, blue']})

I managed to create a for loop that correctly checks if a certain colour (from color_list) is present in the colors column. It then creates a new column that evaluates that condition and assigns True or False to each row. The code is the following:

idx = df.columns.get_loc('colors') + 1

for i in range(len(color_list)):
    df.insert(loc= idx + i, 
                        column='has ' + color_list[i], 
                        value = (df['colors'].str.contains(color_list[i], case=False, na=False)))

It outputs the following:

  object               colors  has green  ...  has red  has yellow  has black
0    car            red, blue      False  ...     True       False      False
1  plane   red, yellow, black      False  ...     True        True       True
2   tree                black      False  ...    False       False       True
3  house          green, blue       True  ...    False       False      False
4  phone  yellow, green, blue       True  ...    False        True      False

Which is correct, the question is can i pass a condition to the value argument of pd.insert?

I want to replace True values with 'has'+(name of the column)' and False with something like 'doesn't have'+(name of column). can this be done in the same loop?

Thanks,

João Victor · Accepted Answer · 2020-04-20 19:25:19Z

2

According to the documentation, the pandas.DataFrame.insert method does not have this functionality.

You can use pandas.DataFrame.apply to do so.

for color in color_list:
    col_name = 'has ' + color
    df[col_name] = df['colors'].apply(lambda x: f'has {color}' if color in x else f'doesn\'t have {color}')

answered Apr 20, 2020 at 19:25

João Victor

4352 silver badges11 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

jcf Over a year ago

For some reason this works fine in Spyder. But in jupyter notebook it returns an error "TypeError: argument of type 'float' is not iterable"

João Victor Over a year ago

You should be assigning a float to color_list variable somewhere. Make sure that color_list is a list.

sammywemmy · Accepted Answer · 2020-04-20 21:23:12Z

1

not sure if this is what you are after; it looks similar to pulling out a 0s and 1s using pandas str dummies function :

repl = {0:False,1:True}

#get a dataframe of 0s and 1s
#dependent on the presence/absence of words
true_false = (df
              .colors
              .str
              .get_dummies(', ')
              .add_prefix('has ')
               #replace the 0s and 1s with False/True
              .replace(repl))

#combine data back to original dataframe
pd.concat([df,true_false],axis=1)


  object    colors         has black    has blue    has green   has red has yellow
0   car     red, blue          False    True        False       True    False
1   plane   red, yellow, black  True    False       False       True    True
2   tree    black               True    False       False       False   False
3   house   green, blue         False   True        True        False   False
4   phone   yellow, green, blue False   True        True        False   True

answered Apr 20, 2020 at 21:23

sammywemmy

28.9k4 gold badges21 silver badges35 bronze badges

1 Comment

jcf Over a year ago

Not exactly what i want, but using a dictionary with. replace gave me added flexibility. Thanks

Collectives™ on Stack Overflow

Inserting New Columns and respective values (on condition) in a Pandas DataFrame

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related