group_by and case_when() function for multiple conditions-CodePudding

I'm struggling with a problem in R. I want to create a new variable (qc) by group_by the variable (NAME and PLOT) using case_when for where "EH” > “PH” then give me B else give me Q......

I have a data set like this:

  df <- tibble(
    NAMEOFEXPERIMENT= c("A","A","A","A","A","A","A","B","B","B","B","B","B","B","B"),
    PLOT= c(2,1,2,1,2,1,2,1,2,1,2,1,2,1,2),
    trait= c("EH","NP","NP","PH","PH","PL","PL","EH","EH","NP","NP","PH","PH","PL","PL"),
    traitValue= c(125,36,36,240,"NA",36,36,90,110,35,33,215,190,36,31)
    )   

 # A tibble: 15 x 4
  NAME  PLOT trait traitValue
  <chr>            <dbl> <chr> <chr>     
  1 A                    2 EH    250       
  2 A                    1 NP    36        
  3 A                    2 NP    36        
  4 A                    1 PH    240       
  5 A                    2 PH    200        
  6 A                    1 PL    36        
  7 A                    2 PL    36        
  8 B                    1 EH    90        
  9 B                    2 EH    110       
 10 B                    1 NP    35        
 11 B                    2 NP    33        
 12 B                    1 PH    215       
 13 B                    2 PH    190       
 14 B                    1 PL    36        
 15 B                    2 PL    31

This is what I want to achieve: If “EH” > “PH” then give me B else give me Q If “PL” > “NP” then give me B else give me Q

Thus, line qc line 4 to be empty since there is no NAME "A", PLOT 1, Trait "EH" to compare with

   # A tibble: 15 x 4
   NAME  PLOT trait traitValue dc
    <chr>            <dbl> <chr> <chr>     <chr>
  1 A                    2 EH    250       B
  2 A                    1 NP    36        Q
  3 A                    2 NP    36        Q
  4 A                    1 PH    240       
  5 A                    2 PH    200       B
  6 A                    1 PL    36        Q
  7 A                    2 PL    36        Q
  8 B                    1 EH    90        Q
  9 B                    2 EH    110       Q
 10 B                    1 NP    35        B
 11 B                    2 NP    33        Q
 12 B                    1 PH    215       Q
 13 B                    2 PH    190       Q
 14 B                    1 PL    36        B
 15 B                    2 PL    31        Q

When I run this code

 dt2 <- df %>%
   group_by(NAME, PLOT) %>%
           traitValue[trait == "EH"] > traitValue[trait == "PH"] ~ "B",
           traitValue[trait == "EH"] < traitValue[trait == "PH"] ~ "Q",
           traitValue[trait == "PL"] > traitValue[trait == "NP"] ~ "B",
           traitValue[trait == "PL"] < traitValue[trait == "NP"] ~ "Q"
           ))

I got this Error

 Error in `mutate()`:
 ! Problem while computing `data_qc = case_when(...)`.
  i The error occurred in group 1: NAME = "A", PLOT = 1.
 Caused by error in`case_when()`: 
 ! `traitValue[trait == "EH"] > traitValue[trait == "PH"] ~ "B"`, traitValue[trait == "EH"] < traitValue[trait == "PH"] ~ "Q"`
 must be length 3 or one, not 0.

CodePudding user response：

I don't fully understand your constraints. You did not specify what would happen if "PH" > "EH" and "PL" > "NP" at the same time. In this case, will the final outcome be "B" or "Q".

However, to get you started I wrote the following code:

## Loading the required libraries
library(dplyr)
library(tidyverse)

## Creating the dataframe
df <- data.frame(
  NAMEOFEXPERIMENT= c("A","A","A","A","A","A","A","B","B","B","B","B","B","B","B"),
  PLOT= c(2,1,2,1,2,1,2,1,2,1,2,1,2,1,2),
  trait= c("EH","NP","NP","PH","PH","PL","PL","EH","EH","NP","NP","PH","PH","PL","PL"),
  traitValue= c(125,36,36,240,200,36,36,90,110,35,33,215,190,36,31)
)  

## Removing duplicates
unique(df)

## Pivot longer to wider
df %>%
  pivot_wider(names_from = trait, values_from = traitValue) %>%
  arrange(NAMEOFEXPERIMENT,PLOT) %>%
  mutate(ConditionalValue1 = ifelse(EH>PH,"B", "Q"),
         ConditionalValue2 = ifelse(PL>NP,"B", "Q"))

Output

# A tibble: 4 x 8
  NAMEOFEXPERIMENT  PLOT    EH    NP    PH    PL ConditionalValue1 ConditionalValue2
  <chr>            <dbl> <dbl> <dbl> <dbl> <dbl> <chr>             <chr>            
1 A                    1    NA    36   240    36 NA                Q                
2 A                    2   125    36   200    36 Q                 Q                
3 B                    1    90    35   215    36 Q                 B                
4 B                    2   110    33   190    31 Q                 Q