Reverse negatives for numeric values according to category in R

user_17755061 0
Viewed 88

I have a data frame that looks like so:

colour evaluation
white 1.00
black - 0.50
white - 0.60
black 1.50

Evaluations is a numeric data type.

I want to create code that flips all the symbols of all evaluations with the colour black, so + becomes - and - becomes +, to create a table like so:

colour evaluation
white 1.00
black 0.50
white - 0.60
black - 1.50

This is the dput() output of the first 20 rows of my data frame:

structure(list(nnue_diff = c(NA, 0.12, -0.05, 0.06, -0.03, 0.13, 
-0.05, 0.02, 0.13, -0.22, -0.08, 0.07, -0.2, 0.24, -0.16, 0.32, 
1.18, -1.23, -0.06, 0), colour = c("white", "black", "white", 
"black", "white", "black", "white", "black", "white", "black", 
"white", "black", "white", "black", "white", "black", "white", 
"black", "white", "black")), row.names = c(NA, -20L), class = c("tbl_df", 
"tbl", "data.frame"))

Does anyone know a method for doing so?

data-wrangling data-cleaning r
2 Answers

Based on updated data

library(dplyr)

df |> mutate(nnue_diff = case_when(colour == "black" ~ nnue_diff * -1 , TRUE ~ nnue_diff))
  • output
# A tibble: 20 × 2
   nnue_diff colour
       <dbl> <chr> 
 1     NA    white 
 2     -0.12 black 
 3     -0.05 white 
 4     -0.06 black 
 5     -0.03 white 
 6     -0.13 black 
 7     -0.05 white 
 8     -0.02 black 
 9      0.13 white 
10      0.22 black 
11     -0.08 white 
12     -0.07 black 
13     -0.2  white 
14     -0.24 black 
15     -0.16 white 
16     -0.32 black 
17      1.18 white 
18      1.23 black 
19     -0.06 white 
20      0    black
user_191983370

Use the swap function from the textclean package

df <- data.frame(colour = c("white", "black", "white", "black"), evaluation = c("+ 1.00", "- 0.50", "- 0.60", "+ 1.50"))

library(dplyr); library(textclean)
df %>% 
  mutate(evaluation = ifelse(colour == 'black', swap(evaluation,'+', '-'), evaluation))

  colour evaluation
1  white     + 1.00
2  black     + 0.50
3  white     - 0.60
4  black     - 1.50
user_88066490
Related