[FEATURE-REQUEST] How to apply same MinMaxScaling to a set of columns?

[yohplala]

Description
Please, how to consider in a MinMaxScaler the set of values coming from several columns (not only one).
Excepted results are illustrated on example below.

import pandas as pd
import vaex as vx

df = pd.DataFrame({'val1': [0.5, 1, 2],
                   'val2': [0.8, 1.2, 3.2],
                   'val3': [1.5, 1.9, 4.0]})
vdf = vx.from_pandas(df)

feat = [('val1', 'val2')] # assuming a tuple of columns would be understood as a set of columns to be grouped together.
scaler = vx.ml.MinMaxScaler(features=feat, prefix='scaled_')
scaler.fit(vdf)
scaler.transform(vdf)

# Min value is in 'val1' and max value is in 'val2'.
Out[32]: 
  #    val1    val2    val3    scaled_val1    scaled_val2
  0     0.5     0.8     1.5        0              0.11111
  1     1       1.2     1.9        0.18518        0.259
  2     2       3.2     4          0.55555        1

Thanks for your help!
Bests

[JovanVeljanoski]

What you describe is not how the MinMaxScaler is expected to work, but nothing stops you from doing this yourself.

Find the min and the max of the features you are interested in, and you can just use the formula for MinMax scaling (you can look this up online). You could write your own function for this, something like

def custom_minmax(df, features, fmin, fmax, a=0, b=1):
  for feat in features:
    df["scaled_"+feat] = (b-a)*(df[feat] - fmin)/(fmax - fmin) + a
  return df

[yohplala]

Thanks @JovanVeljanoski, this is perfect!