Mapping column dataframe with list Python

Question

How to create new column and set values which are results of mapping this dataframe with another object for instance list of lists python?

I have pandas dataframe:

{'a': [15,10,11,9,7,8], 'b':['smth','smth','smth','smth', 'smth', 'smth'].

And list of list:

[[15,10], [11], [9,7,8]]

I want create new column in my dataframe, which will be contain 3 big classes like in my list of lists.

I mean, I want to get this:

{'a': [15,10,11,9,7,8], 'b':['smth','smth','smth','smth', 'smth', 'smth',
'new_column': [0,0,1,2,2,2]}

Please show us what you've tried so far; also I'm not sure I understand what you're trying to do, seems like you can just use dataframe['new_column'] = ... from what you've shown so far — casraf
– casraf, Commented Jul 27, 2017 at 11:04

jezrael · Accepted Answer · 2017-07-27 11:32:11Z

3

You can use map by dict created by dict comprehension, values of list has to be unique:

df = pd.DataFrame({'a': [15,10,11,9,7,8], 'b':['smth','smth','smth','smth', 'smth', 'smth']})

L = [[15,10], [11], [9,7,8]]
#https://stackoverflow.com/q/45349225/2901002
d = { v : i for i,vs in enumerate(L) for v in vs}
#alternative solution
#d = {v: i for i in range(len(L)) for v in L[i]}
print (d)
{7: 2, 8: 2, 9: 2, 10: 0, 11: 1, 15: 0}

df['new_column'] = df['a'].map(d)
print (df)
    a     b  new_column
0  15  smth           0
1  10  smth           0
2  11  smth           1
3   9  smth           2
4   7  smth           2
5   8  smth           2

edited Jul 27, 2017 at 11:32

answered Jul 27, 2017 at 11:08

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

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

8 Comments

Olga Over a year ago

Thank you! This helps me =))

jezrael Over a year ago

Unfortunately only one solution can be accepted, another solution was better?

cs95 Over a year ago

You have my vote because I think your solution is clean. +1

jezrael Over a year ago

@cᴏʟᴅsᴘᴇᴇᴅ - Thank you. I think your solution is overcomplicated. And also list comprehension is the best avoid in pandas with some exceptions.

cs95 Over a year ago

Thank you :) Haha I am a pythonista. However, I do know the basics of other languages (c, java, full stack, apex, etc). My top answer is essentially how to split a number into two parts. It's so basic that I am embarrassed it is at the top. I have written hundreds of better answers. xD Started pandas recently as I am going into a data sciences masters course next month. It will help,

|

cs95 · Accepted Answer · 2017-07-27 11:49:39Z

0

You can use np.where in a list comprehension:

In [926]: import itertools

In [927]: l = np.array(list(itertools.zip_longest(*[[15,10], [11], [9,7,8]], fillvalue=0))).T

In [928]: df['new'] = [np.where(l == i)[0][0] for i in df.a.values]

In [929]: df
Out[929]: 
    a     b  new
0  15  smth    0
1  10  smth    0
2  11  smth    1
3   9  smth    2
4   7  smth    2
5   8  smth    2

answered Jul 27, 2017 at 11:49

cs95

406k106 gold badges744 silver badges797 bronze badges

Collectives™ on Stack Overflow

Mapping column dataframe with list Python

2 Answers 2

8 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

8 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related