How to dynamically replace values in pandas.Series?

Question

So I want to conditionally replace values in pandas.DataFrame using mask function, so imagine we have this dataframe:

     Col1           Col2
0    Qeq            10
1    Qeq            20
2    Qwa            30

What I want is to have an ability to replace all values where Col1 == 'Qwa' to f-string referencing OLD value dynamically. So my desired format is: "-{value}" and desired output is:

     Col1           Col2
0    Qeq            10
1    Qeq            20
2    Qwa            -30

Is it possible using .mask function? I've tried following:

# This works but new value in this case is hardcoded
df['Col2'].mask(df['Col1'] == 'Qwa', other='-' + df['Col2'])
# This does not work as it evaluates whole series in a f-string
df['Col2'].mask(df['Col1'] == 'Qwa', other=f'-{df["Col2"]}')

why not try to filter out the column, then use map()/apply() to effect the change you want? — Ade_1
– Ade_1, Commented Feb 16, 2021 at 14:24
What do you mean by dynamically? If you mean Excel style where you change Col1, Col2 automatically changes, then no, it's not supported. — Quang Hoang
– Quang Hoang, Commented Feb 16, 2021 at 14:26
@Ade_1 Is it possible to reference old value in f-string-like and pass it as an argument? — Disciples
– Disciples, Commented Feb 16, 2021 at 14:28
@QuangHoang By "dynamically" i mean an ability to supply any f-string-like format to be replaced in dataframe/series. NEW-VALUE(OLDVALUE:{value}) or any string like this — Disciples
– Disciples, Commented Feb 16, 2021 at 14:29

Quang Hoang · Accepted Answer · 2021-02-16 14:36:25Z

2

How about apply the format function:

df['Col2'].mask(df['Col1'] == 'Qwa', other=df['Col2'].apply('-{}'.format))

Output:

0     10
1     20
2    -30
Name: Col2, dtype: object

Or

df['Col2'].mask(df['Col1'] == 'Qwa', other=df['Col2'].apply('*{}'.format))

Output:

0     10
1     20
2    *30
Name: Col2, dtype: object

answered Feb 16, 2021 at 14:36

Quang Hoang

151k11 gold badges64 silver badges86 bronze badges

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

Comments

Mayank Porwal · Accepted Answer · 2021-02-16 14:33:43Z

1

Imagine this df:

In [63]: df
Out[63]: 
  Col1  Col2
0  Qeq    10
1  Qeq    20
2  Qwa    30
3  Qwa    40

You can something like, using df.loc:

In [66]: df.loc[df.Col1.eq('Qwa'), 'Col2'] = df.loc[df.Col1.eq('Qwa'), 'Col2'].mul(-1)

In [67]: df
Out[67]: 
  Col1  Col2
0  Qeq    10
1  Qeq    20
2  Qwa   -30
3  Qwa   -40

answered Feb 16, 2021 at 14:33

Mayank Porwal

34.2k9 gold badges45 silver badges65 bronze badges

1 Comment

Disciples Over a year ago

That's nice! Although, I need to work with strings and replace them via pattern using old values

Collectives™ on Stack Overflow

How to dynamically replace values in pandas.Series?

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related