How to use loc on to filter on multiple values

Question

I want to use loc to filter on multiple cities for the column Name

I've tried filtering using several column names:

popdemo_df.loc[popdemo_df['Name'] == 'Richmond city', ['Name'] == Landsdowne']

I get the following error:

False, unexpected EOF while parsing

`popdemo_df.loc[(popdemo_df['Name'] == 'Richmond city') | (popdemo_df['Name'] == 'Landsdowne')]` — Brian, Sep 09 '19 at 17:46

score 3 · Answer 1 · answered Sep 09 '19 at 17:48

In order to combine Boolean indices, you need to surround them with parentheses and use the bitwise operators &, |, or ~, like so:

# Selects rows where either condition is met
popdemo_df.loc[(popdemo_df['Name'] == 'Richmond city') | (popdemo_df['Name'] == 'Landsdowne')]

You can do the same with Series.isin() as well:

popdemo_df.loc[popdemo_df['Name'].isin(['Richmond city', 'Landsdowne'])]

rje · Answer 2 · 2019-09-09T17:49:16.753

0

Use isin:

popdemo_df.loc[popdemo_df['Name'].isin(['Richmond city','Landsdowne'])]

edited Sep 09 '19 at 17:49

answered Sep 09 '19 at 17:46

rje

6,388
1
21
40

While this is the way to do this, I just want to point out to OP that your problem wasn't not using `isin`, your problem was using a comma between dataframe masks and not an pipe (|). As well as calling `['Name']` without a dataframe. – Brian Sep 09 '19 at 17:48
you mean |? it would be impossible for ['Name'] to be two unique values at once – Derek Eden Sep 09 '19 at 17:51

score 0 · Answer 3 · answered Sep 09 '19 at 17:50

Commas are used to separate rows and columns inside df.loc brackets. To combine conditionals, you can use & for and, | for or and ~ for not.

Try this instead:

popdemo_df.loc[:, popdemo_df['Name'] == 'Richmond city' | popdemo_df['Name'] == 'Landsdowne']

How to use loc on to filter on multiple values

3 Answers3