Keep records before and after a specific timestamp

Question

Having a data frame which provides a specific timestamp

dframe1 <- structure(list(id = c(1L, 1L, 1L, 2L, 2L), name = c(\"Google\", 
\"Yahoo\", \"Amazon\",         


        

Answer 1

1. Convert date strings to actual dates.

library(dplyr)

dframe1 <- mutate(dframe1, date = as.Date(date))
dframe2 <- mutate(dframe2, date = as.Date(date))

2. Collapse values of text_sth inside each id, date group in dframe2. They'll appear together anyways in the output.

df2 <- 
  dframe2 %>% 
  group_by(id, date) %>% 
  summarise(text_sth = paste(text_sth, collapse = " "))

df2
#> # A tibble: 10 x 3
#> # Groups:   id [2]
#>       id date       text_sth                              
#>    <int> <date>     <chr>                                 
#>  1     1 2008-10-31 another text other                    
#>  2     1 2008-11-01 test text_sth                         
#>  3     1 2008-11-02 another one test text_sth another text
#>  4     1 2008-11-03 other                                 
#>  5     1 2008-11-04 text here                             
#>  6     1 2008-11-05 text here                             
#>  7     2 2008-10-31 etc                                   
#>  8     2 2008-11-01 test text_sth                         
#>  9     2 2008-11-02 text here another text                
#> 10     2 2008-11-03 text here

3. All the rest: join by id, keep only rows where difference between the date from the first df and the date from the second df is either 1 or -1. Depending on the sign, populate the label variable.

left_join(dframe1, df2, by = "id") %>% 
  mutate(date_diff = as.numeric(date.y - date.x)) %>%
  filter(abs(date_diff) == 1) %>% 
  mutate(label = ifelse(date_diff == -1, "before", "after")) %>% 
  select(id, name, label, text_sth)
#>    id   name  label                               text_sth
#> 1   1 Google before                     another text other
#> 2   1 Google  after another one test text_sth another text
#> 3   1  Yahoo before                     another text other
#> 4   1  Yahoo  after another one test text_sth another text
#> 5   1 Amazon before                                  other
#> 6   1 Amazon  after                              text here
#> 7   2 Amazon before                                    etc
#> 8   2 Amazon  after                 text here another text
#> 9   2 Google before                          test text_sth
#> 10  2 Google  after                              text here