Making a new column in pandas based on conditions of other columns

Question

I would like to make a new column based on an if statement that has conditionals of two or more other columns in a dataframe.

For example, column3 = True if (column1 < 10.0) and (column2 > 0.0).

I have looked around and it seems that other have used the apply method with a lambda function, but i am a bit of a novice on these.

I suppose i could make two additional columns that makes that row a 1 if the condition is met for each column, then sum the columns to check if all conditions are met, but this seems a bit inelegant.

If you provide an answer with apply/lambda, let's suppose the dataframe is called sample_df and the columns are col1, col2, and col3.

Thanks so much!

pansen · Accepted Answer · 2017-04-01 21:03:05Z

2

You can use eval here for short:

# create some dummy data
df = pd.DataFrame(np.random.randint(0, 10, size=(5, 2)), 
                  columns=["col1", "col2"])
print(df)

    col1    col2
0   1       7
1   2       3
2   4       6
3   2       5
4   5       4

df["col3"] = df.eval("col1 < 5 and col2 > 5")
print(df)

    col1    col2    col3
0   1       7       True
1   2       3       False
2   4       6       True
3   2       5       False
4   5       4       False

You can also write it without eval via (df["col1"] < 5) & (df["col2"] > 5).

You may also enhance the example with np.where to explicitly set the values for the positive and negative cases right away:

df["col4"] = np.where(df.eval("col1 < 5 and col2 > 5"), "Positive Value", "Negative Value")
print(df)

    col1    col2    col3    col4
0   1       7       True    Positive Value
1   2       3       False   Negative Value
2   4       6       True    Positive Value
3   2       5       False   Negative Value
4   5       4       False   Negative Value

edited Apr 1, 2017 at 21:03

answered Apr 1, 2017 at 20:52

pansen

6,7034 gold badges21 silver badges33 bronze badges

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

2 Comments

nickm Over a year ago

Thank you, I used the numpy "where" method. Although it seems to not like the "and" keyword, it only works for "&" and "|". Is there a way to assign a value using pandas and not numpy? I see that it returns a list of bools. Would you have to use that as a mask or something? Looking for something like , "If col1 and col2 meet certain conditions, col3 = col1/col2, else none"

pansen Over a year ago

@nickm Yes, you can use the boolean series as mask for whatever values you need. There is also a pandas where which is slightly different.

Collectives™ on Stack Overflow

Making a new column in pandas based on conditions of other columns

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related