Is there a way to manipulate different df columns with different expressions in one line via lambda-CodePudding

I've been experimenting with using lambda through the .apply() method.

I noticed that you can manipulate individual values in each row, and through different expressions, as with this df and manipulation:

test = pd.DataFrame({'credit_score': [608, 675, 502, 699, 850], 'age': [42, 41, 42, 39, 43]}, index=['i1','i2','i3','i4','i5'])

result1 = test.apply(lambda x: (x.i1 *2, x.i2 * 3))

result1

which outputs:

But I haven't been able to do something similar across columns instead of rows.

For example, I thought I might be able to do so with:

result2 = test.apply(lambda x: (x.credit_score *2, x.age * 3), axis=1)
result2

But this approach outputs tuples in the non-index column:

Is there a way to preserve the columns and modify the values vertically downward from the columns, instead of horizontally from rows? Basically I'm trying to understand how it might be possible to do more comprehensive calculations from a single code line—though I'd also be interested in other workarounds.

CodePudding user response：

You are doing two different operations in your example...

In the first case you are applying on the first and second rows only lambda x: (x.i1 *2, x.i2 * 3). i1 and i2 are index of the rows

Second operation you are applying to the first and second columns (NOT rows) hence you get two rows of output

I assume you were looking for something like this?

credit_score age    new_column
i1  608       42    1216
i2  675       41    1350
i3  502       42    1004
i4  699       39    1398
i5  850       43    1700

This will give you the above ouput test['new_column']=test.apply(lambda x: x.credit_score *2, axis=1) Updated: The OP was looking for apply to return results to multiple columns test[["ex1","ex2"]] = test.apply(lambda x: [x.credit_score *2, x.age * 3], axis=1, result_type="expand")