How to select the max value of each row (not all columns) and mutate 2 columns which are the max val-CodePudding

This is the original dataframe. I want to know the max value between 'a' and 'b', and mutate max value and variable name as two columns.

df <- data.frame(lon = c(102,103,104,105),
                 lat = c(31,32,33,34),
                 a = c(4,3,7,6),
                 b = c(5,2,4,9))

The target dataframe is like this.

dftarget <- data.frame(lon = c(102,103,104,105),
                       lat = c(31,32,33,34),
                       a = c(4,3,7,6),
                       b = c(5,2,4,9),
                       max = c(5,3,7,9),
                       type = c('b','a','a','b'))

CodePudding user response：

Method 1

Simply use pmax and max.col function to identify the maximum values and columns.

library(dplyr)

df %>% mutate(max = pmax(a,b), type = colnames(df)[max.col(df[,3:4])   2 ])

Method 2

Or first re-shape your data to a "long" format for easier manipulation. Then use mutate to extract max values and names. Finally change it back to a "wide" format and relocate columns according to your target.

df %>% 
  pivot_longer(a:b, names_to = "colname") %>% 
  group_by(lon, lat) %>% 
  mutate(max = max(value), 
         type = colname[which.max(value)]) %>% 
  pivot_wider(everything(), names_from = "colname", values_from = "value") %>% 
  relocate(max, type, .after = b)

Output

# A tibble: 4 × 6
# Groups:   lon, lat [4]
    lon   lat     a     b   max type 
  <dbl> <dbl> <dbl> <dbl> <dbl> <chr>
1   102    31     4     5     5 b    
2   103    32     3     2     3 a    
3   104    33     7     4     7 a    
4   105    34     6     9     9 b