CODEWITHSUNDEEP
pythonpandas
s5s
s5s

Reputation: 12204

pandas multiply values by scaling factor in another DataFrame

I have some data with an id and val columns in DataFrame df and some scaling factors in DataFrame scaling such that it has a scaling factor for each value in id

df = pd.DataFrame(data=dict(id=['a', 'a', 'a', 'b', 'b', 'c'], val=[1, 2, 3, 10, 11, 100]))
scaling = pd.DataFrame(data=dict(id=['a', 'b', 'c'], scaling=[1, 0.1, 0.01]))

These look like this:

In[23]: df
Out[24]: 
  id  val
0  a    1
1  a    2
2  a    3
3  b   10
4  b   11
5  c  100
scaling
Out[25]: 
  id  scaling
0  a     1.00
1  b     0.10
2  c     0.01

I now want to multiply the data in df by the scaling factors. I can do it this way but this feels awkward and is possibly inefficient?

df['val'] = df['val'] * df.merge(scaling, left_on='id', right_on='id')['scaling']

Is there a better way of applying the factors?

Upvotes: 2

Views: 707

Answers (2)

Joe
Joe

Reputation: 889

You can map by creating a dictionary from the columns in your scaling dtaframe.

>>> df = pd.DataFrame(data=dict(id=['a', 'a', 'a', 'b', 'b', 'c'], val=[1, 2, 3, 10, 11, 100]))

>>> scaling = pd.DataFrame(data=dict(id=['a', 'b', 'c'], scaling=[1, 0.1, 0.01]))

>>> scaling_dict = dict(zip(scaling['id'], scaling['scaling']))

>>> df['multiplier'] = df['id'].map(scaling_dict)

>>> df['val'] = df['val']*df['multiplier']

>>> df
  id  val  multiplier
0  a  1.0        1.00
1  a  2.0        1.00
2  a  3.0        1.00
3  b  1.0        0.10
4  b  1.1        0.10
5  c  1.0        0.01

Upvotes: 0

Quang Hoang
Quang Hoang

Reputation: 150815

You can use map with set_index on scaling:

df['val'] * df['id'].map(scaling.set_index('id').scaling)

Output:

0    1.0
1    2.0
2    3.0
3    1.0
4    1.1
5    1.0
dtype: float64

Upvotes: 2

Related Questions