Pandas Get a list of index from dataframe.loc

Question

I have looked through various sites and SO posts.Seems easy but somehow i am stuck with this.I am using

print frame.loc[(frame['RR'].str.contains("^[^123]", na=False)), 'RR'].isin(series1.str.slice(1))

to get

3     True
4    False
8    False
Name: RR, dtype: bool

Now,somehow i want the indexes only so that i can use that in dataframe.drop. Basically all the indexes where value is True , i have to grab indexes and drop them.Is there any other way as well without using indexes?

score 1 · Accepted Answer · 2016-10-15 14:34:18Z

1

You are testing two conditions on the same column so these can be combined (and negated):

frame[~((frame['RR'].str.contains("^[^123]", na=False)) & (frame['RR'].isin(series1.str.slice(1))))]

Here, after ~ operator, it checks whether a particular row satisfies both conditions - same as the boolean array you get in the end. With ~, you turn True's to False's and False's to True's. Finally, frame[condition] returns the rows that satisfy the final condition with boolean indexing.

In a more readable format:

condition1 = frame['RR'].str.contains("^[^123]", na=False)
condition2 = frame['RR'].isin(series1.str.slice(1))
frame[~(condition1 & condition2)]

As an alternative (requires 0.18.0), you can get the indices of the True elements with:

frame.loc[(frame['RR'].str.contains("^[^123]", na=False)), 'RR'].isin(series1.str.slice(1))[lambda df: df].index

edited Oct 15, 2016 at 14:34

answered Oct 15, 2016 at 14:29

user2285236

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

4 Comments

vks Over a year ago

So which of the two would be faster considering frames would have a million entries

user2285236 Over a year ago

Most probably the first one. The second one first returns another frame and tests another condition just to give the indices - which then need to be used in another method (drop). I've tested on a small dataframe and the first one is 2x faster (without considering the drop part). Of course this may change based on the entries you have in series1, in frame['RR'], etc.

vks Over a year ago

Cool!!! thanx a lot!!! can you share some good resource on pandas? the info is all splattered across....

user2285236 Over a year ago

You are welcome. My main source are Q&A's here and the docs. There is also Wes McKinney's book (although a little dates, a new edition is about to come). Lastly, Tom Augspurger's Modern Pandas series is quite nice.

Collectives™ on Stack Overflow

Pandas Get a list of index from dataframe.loc

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related