Ytukyg

I have two data frames (tibbles) with 2 variables each:

• df.POS: ID (ID variable); DATE (Date of positive lab test)


• df.NEG: ID (ID variable); data (Date of negative lab tests (more than 1 test).

Please note that data is a list variable, created with the nest() function of the tidyr package.

library(tidyverse)

library(lubridate)



# negative tests

dates.neg <- ymd(c('2018-02-01', '2018-02-06', '2018-02-10', 

             '2018-02-21', '2018-04-05'))

df.NEG <- tibble(ID = paste0('ID_', rep(1, 5)),

          DATE = dates.neg) %>%

       group_by(ID) %>% 

          nest()

df.NEG



## # A tibble: 1 x 2

##   ID    data            

##   <chr> <list>          

## 1 ID_1  <tibble [5 × 1]>





dates.pos <- ymd(c('2018-02-07', '2018-02-12', '2018-02-13', 

             '2018-02-20', '2018-02-21', '2018-03-18'))



df.POS <- tibble(ID = paste0('ID_', rep(1, 6)),

           DATE = dates.pos)

df.POS



## # A tibble: 6 x 2

##   ID    DATE      

##   <chr> <date>    

## 1 ID_1  2018-02-07

## 2 ID_1  2018-02-12

## 3 ID_1  2018-02-13

## 4 ID_1  2018-02-20

## 5 ID_1  2018-02-21

## 6 ID_1  2018-03-18

I would like to find out for which of the positive tests there was also a negative test up to 2 days after the positive test result. I've tried using the map2() function of the purrr package

df.TOTAL <- df.POS %>%

  left_join(df.NEG, by = 'ID') %>%

    mutate(TIME = interval(DATE, DATE + days(2)),

           RESULT = map2(data, "DATE", TIME, ~ .x %within% .y)) 

Unfortunaltely, my code doesn't work. The RESULT variable should be logical and return TRUE in case of a negative test result up to 2 days after the positive test. Instead it is a list and returns NULL.

df.TOTAL



## # A tibble: 6 x 5

##   ID    DATE       data             TIME                           RESULT

##   <chr> <date>     <list>           <S4: Interval>                 <list>

## 1 ID_1  2018-02-07 <tibble [5 × 1]> 2018-02-07 UTC--2018-02-09 UTC <NULL>

## 2 ID_1  2018-02-12 <tibble [5 × 1]> 2018-02-12 UTC--2018-02-14 UTC <NULL>

## 3 ID_1  2018-02-13 <tibble [5 × 1]> 2018-02-13 UTC--2018-02-15 UTC <NULL>

## 4 ID_1  2018-02-20 <tibble [5 × 1]> 2018-02-20 UTC--2018-02-22 UTC <NULL>

## 5 ID_1  2018-02-21 <tibble [5 × 1]> 2018-02-21 UTC--2018-02-23 UTC <NULL>

## 6 ID_1  2018-03-18 <tibble [5 × 1]> 2018-03-18 UTC--2018-03-20 UTC <NULL>

Can anyone help?

I would appreciate some help. Thanks very much in advance!

First, note that you can test whether any element from a vector of "negative" dates falls within the "positive" interval like so:

any(dates.neg %within% interval(dates.pos[1], dates.pos[1] + days(2)))

# [1] FALSE

This suggests the following approach using map2 -- or more usefully, map2_lgl:

df.TOTAL <- df.POS %>%

  left_join(df.NEG, by = 'ID') %>%

    mutate(TIME = interval(DATE, DATE + days(2)),

           RESULT = map2_lgl(data, TIME, ~any(.x$DATE %within% .y)))

# # A tibble: 6 x 5

#   ID    DATE       data             TIME                           RESULT

#   <chr> <date>     <list>           <S4: Interval>                 <lgl> 

# 1 ID_1  2018-02-07 <tibble [5 x 1]> 2018-02-07 UTC--2018-02-09 UTC FALSE 

# 2 ID_1  2018-02-12 <tibble [5 x 1]> 2018-02-12 UTC--2018-02-14 UTC FALSE 

# 3 ID_1  2018-02-13 <tibble [5 x 1]> 2018-02-13 UTC--2018-02-15 UTC FALSE 

# 4 ID_1  2018-02-20 <tibble [5 x 1]> 2018-02-20 UTC--2018-02-22 UTC TRUE  

# 5 ID_1  2018-02-21 <tibble [5 x 1]> 2018-02-21 UTC--2018-02-23 UTC TRUE  

# 6 ID_1  2018-03-18 <tibble [5 x 1]> 2018-03-18 UTC--2018-03-20 UTC FALSE 

Thanks to @ubutun for improving the answer.