Filtering based on information from 3 columns-CodePudding

For some context, I wanna have a way to filter for a certain value in a certain test. If a person has the required value, all his tests and their results should remain in the data. If not, all their data should be removed. The following code shoul help.

person <- c('pers1', 'pers1', 'pers2', 'pers2', 'pers2', 'pers3', 'pers3', 'pers4', 'pers5', 'pers5', 'pers6') 
test <-c('a', 'b', 'a', 'b', 'c', 'b', 'c', 'a' , 'a', 'c', 'b' ) 
value <-c(2, 3, 4, 2, 1, 5, 7, 4, 1, 3, 1)

data <- data.frame(person,test,value)
head(data,20)

With the following code, I'm removing all the people who haven't had an "a" test. Which are person 3 and 6. I'm keeping every person who have had a test "a" done, with all their other test's stored aswell, so I later on can do some statistics and correlations.

data1 <- data[data$person %in% data[data$test=='a',]$person,]
data1

However, I want to add another layer on the "filtering". I want to filter out the people who's value in test "a" are 3 or above. Which means I would only have Person 1 and 5 left (with their other tests aswell.) to make things clear, this is what I want to have left:

person1 <- c('pers1', 'pers1', 'pers5', 'pers5')
test1 <- c('a', 'b', 'a', 'c')
value1 <- c(2,3,1,3)

data1 <- data.frame(person1,test1,value1)
data1

I hope this is enough data for you to work with. First time here posting a code.

CodePudding user response：

You can add that requirement to your sub-query

data[data$person %in% data[data$test=='a' & data$value<3,]$person,]
#    person test value
# 1   pers1    a     2
# 2   pers1    b     3
# 9   pers5    a     1
# 10  pers5    c     3

If you wanted to use dplyr, you could use group_by() and filter()

library(dplyr)
data %>% 
  group_by(person) %>% 
  filter(any(test=="a" & value < 3))

CodePudding user response：

Using ave from base R

subset(data, ave(test == 'a' & value < 3, person, FUN = any))
   person test value
1   pers1    a     2
2   pers1    b     3
9   pers5    a     1
10  pers5    c     3