Filter of dplyr (with any of rows)

Question

I'm new in the R world, these days I have problem with the filter function of dplyr.I have a df and a I need to filter the values greater than a number, but in some rows I have multiples values (separeted by; ). For example I have this df

ID   value1    value2
1    1;0;3.4    4
2    3          5
3    0.5;2;1.3  0
4    5;0.1      3

My filter is if one or more in value1 is greater or equal than 3. I use the function filter of dplyr because I need realize others filters to the df. This is my code.

filt <- df %>% filter(any(as.numeric(unlist(strsplit(value1,';',fixed=TRUE))) >=3))

But in this case the any function considers all the values of the df (not by row, as I thought), so, I obtein all df and that is not correct.

I need obtain row with id 1,2 and 4 for this example.

I think I should check by row and separate by; , but I don't know how to do this with the filter function of dplyr.

Thanks so much!

Answer 1

An option would be to split the 'value1' with separate_rows from tidyr , grouped by 'ID', filter groups having any element in 'value1' greater than or equal to 3, then summarise the columns by paste ing and getting the first element of 'value2'

library(dplyr)
library(tidyr)
df %>%
    separate_rows(value1, sep = ";", convert = TRUE) %>%
    group_by(ID) %>%
    filter(any(value1 >=3)) %>%
    summarise(value1 = str_c(value1, collapse=";"), value2 = first(value2))
# A tibble: 3 x 3
#     ID value1  value2
#  <int> <chr>    <int>
#1     1 1;0;3.4      4
#2     2 3            5
#3     4 5;0.1        3

---

Or using map with strsplit

library(purrr)
df %>% 
   filter(map_lgl(strsplit(value1, ";"), ~ any(as.numeric(.x) >=3)))
#   ID  value1 value2
#1  1 1;0;3.4      4
#2  2       3      5
#3  4   5;0.1      3

data

df <- structure(list(ID = 1:4, value1 = c("1;0;3.4", "3", "0.5;2;1.3", 
"5;0.1"), value2 = c(4L, 5L, 0L, 3L)), class = "data.frame", row.names = c(NA, 
-4L))