I have a dataset similar to this:

> dput(df1)
structure(list(Person_Id = c(123L, 123L, 123L, 123L, 123L, 
                             567L, 567L, 567L, 888L, 888L), 
               Result = c("Yes", "Yes", "Yes", "Yes", "Yes", 
                          "No", "No", "No", "No", "Yes")), 
          class = "data.frame", row.names = c(NA, -10L))

As you can see, the Result values for Person_id 123 and 567 are consistent throughout the entire dataframe. However, the Result values for person 888 differs. I want to filter this dataset:

1. If the result differs for a given Person_id, I would like them to be removed.
2. Keep only one row for a given Person_id IF the Result is consistent throughout the dataframe.

The output dataframe should look like:

> dput(df2)
structure(list(Person_Id = c(123L, 567L), 
               Result = c("Yes", "No")), 
          class = "data.frame", row.names = c(NA, -2L))

CodePudding user response：

A base R approach using aggregate

na.omit(aggregate(Result ~ Person_Id, df1, function(x) 
  ifelse(length(unique(x)) == 1, unique(x), NA)))
  Person_Id Result
1       123    Yes
2       567     No

CodePudding user response：

Here's a dplyr solution:

library(dplyr)

df1 %>% 
  group_by(Person_Id) %>% 
  filter(n_distinct(Result) == 1) %>% 
  unique %>% ungroup

#> # A tibble: 2 x 2
#>   Person_Id Result
#>       <int> <chr> 
#> 1       123 Yes   
#> 2       567 No