rescale the selective range into 0 - 1?

Question

I have a data frame like this:

df <- data.frame(FC=c(1, 2, 4, 2, 1, -4, -2))

And I'm trying to rescale the positive values between 0-1 and negative values also between 0- -1 so the output would be:

 FC
0.25
 0.5
   1
 0.5
0.25
  -1
-0.5

I have tried "scales" package in r

df$FC <- rescale(df$FC, to = c(0, 1))

But it is rescaling all values between 0-1.

How can I normalize columns by their individual range?

I really appreciate any help you can provide.

Answer 1

Here is a base R option

transform(
    df,
    FC_scaled = FC / abs(ifelse(FC > 0, max(FC), min(FC)))
)

which gives

  FC FC_scaled
1  1      0.25
2  2      0.50
3  4      1.00
4  2      0.50
5  1      0.25
6 -4     -1.00
7 -2     -0.50

Answer 2

You could just slice your data into positives and negatives and handle each individually.

df <- data.frame(FC=c(1, 2, 4, 2, 1, -4, -2))

library(scales)

df[which(df$FC>0),] <- rescale(df[which(df$FC>0),], to=c(.25,1))
df[which(df$FC<0),] <- rescale(df[which(df$FC<0),], to=c(-.5,-1))

df

Output:

     FC
1  0.25
2  0.50
3  1.00
4  0.50
5  0.25
6 -0.50
7 -1.00

Please note however, that I did not specify the range to be between 0 and 1 because that would not get you the output you were looking for. If I did the output would be this:

          FC
1  0.0000000
2  0.3333333
3  1.0000000
4  0.3333333
5  0.0000000
6  0.0000000
7 -1.0000000

Answer 3

We may also use sign with similar logic as in @ThomasIsCoding post

df$FC_scaled <- with(df, FC/setNames(range(FC),
           c(-1, 1))[as.character(sign(FC))] * sign(FC))

-output

 df
  FC FC_scaled
1  1      0.25
2  2      0.50
3  4      1.00
4  2      0.50
5  1      0.25
6 -4     -1.00
7 -2     -0.50