Add missing intervals to incomplete start-stop data

Some start-stop datasets may be missing some intervals for some cases. This function adds those intervals to the data, which might be needed for further processing.

Usage

fill_gaps_start_stop(data, id, start="start", stop="stop",
                     first_time=NULL, last_time=NULL,
                     missing_indicator=TRUE, ...)

Arguments

data: A data.table like object including at least three columns: id (the unique case identifier), start (the beginning of the time-interval) and stop (the end of the time-interval). May also be any object that can be coerced to be a data.table, such as a data.frame or a tibble. Intervals should be coded as (start, stop].
id: A single character string specifying a column in id specifying the unique case identifier.
start: A single character string specifying a column in data specifying the beginning of a time-interval. Defaults to "start".
stop: A single character string specifying a column in data specifying the ending of a time-interval. Defaults to "stop".
first_time: Corresponds to the argument of the same name in the merge_start_stop function.
last_time: Corresponds to the argument of the same name in the merge_start_stop function.
missing_indicator: Either TRUE (default) or FALSE, specifying whether a logical column should be added to the data which includes an indicator whether the corresponding interval was present in the original data or not. If TRUE, this column will be added using the name ".in_data". This column is TRUE if the interval was present and FALSE if it was not.
...: Further arguments passed to the merge_start_stop function.

Details

If the input data does not have any missing intervals, the input stays unchanged. If missing intervals are added, they have a value of NA in all columns other than start, stop and id.

Internally, this function simply calls the merge_start_stop function with appropriate input argument (merging the input data with an essentially empty second dataset using all=TRUE). As such, all arguments of the merge_start_stop function, such as setting default values using the defaults argument may also be used here.

Value

Returns a single data.table containing the start-stop data with all missing intervals added.

Author

Robin Denz

Examples

library(MatchTime)
library(data.table)

## define some example data
# this dataset is missing two intervals: (24, 28] for id = 1 and
# (245, 343] for id = 2
data <- data.table(id=c(1, 1, 1, 2, 2, 3),
                   start=c(0, 28, 66, 25, 343, 10),
                   stop=c(24, 66, 143, 245, 1233, 3214),
                   A=c(10, 122, 3434, 223, 233, 0.46))

# add those intervals back
out <- fill_gaps_start_stop(data, id="id")
print(out)
#> Key: <id, start>
#>       id start  stop .in_data       A
#>    <num> <num> <num>   <lgcl>   <num>
#> 1:     1     0    24     TRUE   10.00
#> 2:     1    24    28    FALSE      NA
#> 3:     1    28    66     TRUE  122.00
#> 4:     1    66   143     TRUE 3434.00
#> 5:     2    25   245     TRUE  223.00
#> 6:     2   245   343    FALSE      NA
#> 7:     2   343  1233     TRUE  233.00
#> 8:     3    10  3214     TRUE    0.46

# add the missing intervals, but also add missing intervals from 0 to first
# observed value per id
out <- fill_gaps_start_stop(data, id="id", first_time=0)
print(out)
#> Key: <id, start>
#>        id start  stop .in_data       A
#>     <num> <num> <num>   <lgcl>   <num>
#>  1:     1     0    24     TRUE   10.00
#>  2:     1    24    28    FALSE      NA
#>  3:     1    28    66     TRUE  122.00
#>  4:     1    66   143     TRUE 3434.00
#>  5:     2     0    25    FALSE      NA
#>  6:     2    25   245     TRUE  223.00
#>  7:     2   245   343    FALSE      NA
#>  8:     2   343  1233     TRUE  233.00
#>  9:     3     0    10    FALSE      NA
#> 10:     3    10  3214     TRUE    0.46