Specialized "long to wide" reshaping

On the surface, conjecture() appears similar to tidyr::pivot_wider(), but uses different logic tailored to a specific type of dataset:

  • column corresponding to names_from contains only 2 levels

  • there is no determinate combination of elements to fill 2 columns per row.

See vignette("conjecture") for more details.

conjecture(data, sort_by, names_from, names_first)

Arguments

data

A data frame to reshape.
sort_by

Column name, as symbol. Plays a similar role as values_from in pivot_wider(), but also serves as sorting dimension for underlying conjecture algorithm.
names_from

Column name, as symbol. Used to differentiate anterior/posterior observations. Column must only contain 2 levels (missing values not allowed).
names_first

level in variable specified by names_from indicating anterior observation.

Value

An object of the same type as data.

Details

conjecture() uses the following routine to match elements:

  1. Values in sort_by are separated into two vectors: anterior and posterior.

  2. Each anterior element is matched with the closest posterior element measured by sort_by.

Examples

# See vignette("conjecture") for more examples

conjecture(comms, timestamp, type, "send")

#> # A tibble: 24,958 x 4
#>    station msg_code send                receive            
#>    <chr>      <int> <dttm>              <dttm>             
#>  1 D           2537 1999-01-21 03:37:57 1999-02-15 21:56:29
#>  2 C            720 1999-01-24 11:52:07 1999-02-22 12:24:57
#>  3 D           2531 1999-01-26 11:50:58 1999-02-09 07:14:33
#>  4 D           2992 1999-01-27 02:48:56 1999-02-22 12:05:55
#>  5 A           2262 1999-01-27 09:19:56 1999-02-11 19:43:42
#>  6 B           1785 1999-01-27 12:11:04 1999-02-07 03:07:50
#>  7 C           1624 1999-01-27 15:33:09 1999-02-20 05:07:54
#>  8 C           2280 1999-01-27 20:06:18 1999-02-18 11:25:13
#>  9 B           1170 1999-01-28 00:55:33 NA                 
#> 10 B           2137 1999-01-28 02:30:30 1999-02-26 17:13:21
#> # … with 24,948 more rows