remove rows with conditions based on columns-CodePudding

I am trying to remove rows that belong to the same TRIAL group and Trait group with empty traitvalue

dt <- tibble(
TRIAL = c("M20-07RND1", "M20-07RND1", "M20-07RND1", " M20-07RND1", " M20-06RND5", " M20-06RND5", " M20-06RND5", " M20-06RND5", " V-01RND1", " V-01RND1", " V-01RND1", "V-01RND1"),
  PLOT= c(1, 2, 3, 1, 6, 3, 2, 3, 1,67,34,34),
  trait = c("ASI", "ASI", "AD", "AD", "ASI", "ASI", "FY", "FY", "FOR", "FOR", "CY", "CY"),
  trialValue = c(2, 7, 2, 2, NA,34 ,NA ,NA ,10 ,NA ,NA ,NA)
)

# # A tibble: 11 x 4
 #      TRIAL     PLOT  trait    trialValue
        <chr>     <dbl> <chr>      <dbl 
  #1   M20-07RND1   1    ASI        2
  #2   M20-07RND1   2    ASI        7
  #3   M20-07RND1   3    AD         2
  #4   M20-07RND1   1    AD         2
  #5   M20-06RND5   6    ASI               
  #6   M20-06RND5   3    ASI       34
  #7   M20-06RND5   2    FY
  #8   M20-06RND5   3    FY
  #9   V-01RND1     1    FOR       10
  #10  V-01RND1     67   FOR
  #11  v-01RND1     34   CY
  #12  v-01RND1     34   CY

This is what I want to achieve: If a group of trait has an empty traitvalue and belong to the same TRIAL Group then remove the whole rows where the condition met.

 # # A tibble: 11 x 4
 #      TRIAL     PLOT  trait     trialValue
        <chr>     <dbl> <chr>       <dbl 
  #1   M20-07RND1   1    ASI         2
  #2   M20-07RND1   2    ASI         7
  #3   M20-07RND1   3    AD          2
  #4   M20-07RND1   1    AD          2
  #5   M20-06RND5   6    ASI               
  #6   M20-06RND5   3    ASI        34
  #9   V-01RND1     1    FOR        10
  #10  V-01RND1     67   FOR

I don’t know how to go about it, kindly help out.

CodePudding user response：

library(dplyr)

dt %>% 
  group_by(TRIAL, trait) %>% 
  filter(sum(trialValue, na.rm = TRUE) != 0)

# A tibble: 8 x 4
# Groups:   TRIAL, trait [5]
  TRIAL          PLOT trait trialValue
  <chr>         <dbl> <chr>      <dbl>
1 "M20-07RND1"      1 ASI            2
2 "M20-07RND1"      2 ASI            7
3 "M20-07RND1"      3 AD             2
4 " M20-07RND1"     1 AD             2
5 " M20-06RND5"     6 ASI           NA
6 " M20-06RND5"     3 ASI           34
7 " V-01RND1"       1 FOR           10
8 " V-01RND1"      67 FOR           NA

CodePudding user response：

Use any:

library(dplyr)
dt %>% 
  group_by(TRIAL, trait) %>% 
  filter(any(!is.na(trialValue)))

output

# A tibble: 8 x 4
# Groups:   TRIAL, trait [5]
  TRIAL          PLOT trait trialValue
  <chr>         <dbl> <chr>      <dbl>
1 "M20-07RND1"      1 ASI            2
2 "M20-07RND1"      2 ASI            7
3 "M20-07RND1"      3 AD             2
4 " M20-07RND1"     1 AD             2
5 " M20-06RND5"     6 ASI           NA
6 " M20-06RND5"     3 ASI           34
7 " V-01RND1"       1 FOR           10
8 " V-01RND1"      67 FOR           NA