Rolling Conditional Pandas DataFrame Column

Question

How would I be able to write a rolling condition apply to a column in pandas?

import pandas as pd
import numpy as np    

lst = np.random.random_integers(low = -10, high = 10, size = 10)
lst2 = np.random.random_integers(low = -10, high = 10, size = 10)

#lst = [ -2  10 -10  -6   4   2  -5   4   9   3]
#lst2 = [-7  5  6 -4  7  1 -4 -6 -1 -4]
df = pandas.DataFrame({'a' : lst, 'b' : lst2})

Given a dataframe, namely 'df', I want to create a column 'C' such that it will display True if the element in a > 0 and b > 0 or False if a < 0 and b < 0.

For the rows that do not meet this condition, I want to roll the entry in the previous row to the current one (i.e. if the previous row has value 'True' but does not meet the specified conditions, it should have value 'True'.)

How can I do this?

Follow-up Question : How would I do this for conditions a > 1 and b > 1 returns True or a < -1 and b < -1 returns False?

cs95 · Accepted Answer · 2018-07-09 15:12:05Z

2

I prefer to do this with a little mathemagic on the signs.

i = np.sign(df.a)
j = np.sign(df.b)

i = i.mask(i != j).ffill()
i >= 0

# for your `lst` and `lst2` input 
0    False
1     True
2     True
3    False
4     True
5     True
6    False
7    False
8    False
9    False
Name: a, dtype: bool

As long as you don't have to worry about integer overflows, this works just fine.

edited Jul 9, 2018 at 15:12

answered Jul 9, 2018 at 14:43

cs95

406k106 gold badges744 silver badges797 bronze badges

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

6 Comments

Ziv Lotzky Over a year ago

That worked for the first entry where a condition was not met, however, it didn't reapply the condition when I tried to use it for the entire dataframe.

cs95 Over a year ago

@ZivLotzky Sorry about that... try it now?

Ziv Lotzky Over a year ago

Not exactly... I want to return True if a > 0 and b > 0 and False if a < 0 and b < 0. However, your code is returning True when either condition is met and False when neither is met. Instead, if a condition is not met, I wish to return the previous result. For example, if row 1 returns False, but no condition is met for row 2, it should also return False.

Ziv Lotzky Over a year ago

As a follow-up question, how would I do this for conditions a > 1 and b > 1 returns True or a < -1 and b < -1 returns False? This might be a different answer since you do not have a uniform condition of greater than or equal to 0? @coldspeed

cs95 Over a year ago

@ZivLotzky Yeah, I think the answer will be different... would you like to open another question?

|

Ziv Lotzky · Accepted Answer · 2018-07-09 15:14:49Z

1

i = np.sign(df.a)
j = np.sign(df.b)

i.mask(i != j).ffill().ge(0)

answered Jul 9, 2018 at 15:14

Ziv Lotzky

458 bronze badges

1 Comment

cs95 Over a year ago

Yes, this is what I came up with in the end... sorry for the retries, brain's a bit slow this monday... refresh your page.

Collectives™ on Stack Overflow

Rolling Conditional Pandas DataFrame Column

2 Answers 2

6 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

6 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related