How to write a lambda function that is conditional on two variables (columns) in python

Question

I have a data set, df, with two variables, x and y. I want to write a function that does the following:

x if x>100 and y<50 else y

I am used to doing data analysis in STATA so I'm relatively new to pandas for data analysis. If it helps, in stata it would look like:

replace x = cond(x>100 & y<50, x, y)

In other words, the function is conditional on two columns in df and will return a value from one variable or the other in each row depending on whether the condition is met.

So far I have been creating new variables through new functions like:

df.dummyVar = df.x.apply(lambda x: 1 if x>100 else 0)

Using StackOverflow and the documentation I have only been able to find how to apply a function dependent on a single variable to more than one column (using the axis option). Please help.

EdChum · Accepted Answer · 2014-07-16 21:16:46Z

14

Use where:

df['dummyVar '] = df['x'].where((df['x'] > 100) & (df['y'] < 50), df['y'])

This will be much faster than performing an apply operation as it is vectorised.

answered Jul 16, 2014 at 21:16

EdChum

397k204 gold badges837 silver badges583 bronze badges

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

1 Comment

seeiespi Over a year ago

This is exactly what I needed. And this is great because I can already see how I can expand it to conditionals on 3 or more variables. Thank you!

James Mills · Accepted Answer · 2014-07-16 21:08:47Z

6

Like this:

f = lambda x, y: x if x>100 and y<50 else y

Lambda(s) in Python are equivalent to a normal function definition.

def f(x, y):
    return x if x>100 and y<50 else y

NB: The body of a Lambda must be a valid expression. This means you cannot use things like: return for example; a Lambda will return the last expression evaluated.

For some good reading see:

edited Jul 16, 2014 at 21:08

answered Jul 16, 2014 at 21:02

James Mills

19.1k4 gold badges53 silver badges63 bronze badges

2 Comments

seeiespi Over a year ago

I had actually written a function like this but was unable to implement so that it would run through each row without a loop. The answer provided by EdChum does exactly that. If you know of a way to accomplish that using this defined function I'm sure I could make use of that in the future. Thank you for you input :)

James Mills Over a year ago

@seeiespi You originally asked for "How to create a lambda function that takes two arguments?" -- This is how :) -- EdChum provided you with an answer that is more aligned with what you're intentions are/were with your dataset(s) and pandas.

seeiespi · Accepted Answer · 2018-04-26 19:51:30Z

2

There's now an pretty easy way to do this. Just use apply on the dataset:

df['dummy'] = df.apply(lambda row: row['x'] if row['x'] > 100 and row['y'] < 50 else row['y'])

answered Apr 26, 2018 at 19:51

seeiespi

3,8642 gold badges37 silver badges39 bronze badges

1 Comment

Manish Singla Over a year ago

I need to write the axis = 1 to make it work

Collectives™ on Stack Overflow

How to write a lambda function that is conditional on two variables (columns) in python

3 Answers 3

1 Comment

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related