Wide <ROOT>_<VISIT> data to long format

One row per subject and visit, one column per root. Derived variables are roots like any other (LOG_NFL_V0 -> root LOG_NFL). Columns are re-typed after reshaping: numeric when possible, otherwise factor (levels of the first visit kept).

Usage

visit_to_long(
  df,
  roots,
  id_col,
  visits = c("V0", "V3"),
  keep = NULL,
  visit_col = "VISIT",
  time_values = NULL
)

Arguments

df

A data frame in wide format.

roots

Root names to reshape.

id_col

Identifier column (name or bare symbol).

visits

Visit suffixes, in chronological order.

keep

Time-invariant columns to keep (names or tidyselect).

visit_col

Name of the visit column created.

time_values

Optional named vector visit -> time (e.g. c(V0 = 0, V3 = 36)), stored in a TIME column.

Value

A tibble.

Examples

long <- visit_to_long(cohort_sim, c("NFL", "MOCA"), id_col = ID, keep = c(GROUP, AGE),
                      time_values = c(V0 = 0, V3 = 36))
head(long)
#> # A tibble: 6 × 7
#>   ID    GROUP     AGE VISIT   NFL  MOCA  TIME
#>   <chr> <fct>   <dbl> <fct> <dbl> <dbl> <dbl>
#> 1 P001  Placebo    80 V0     26.8    24     0
#> 2 P001  Placebo    80 V3     NA      NA    36
#> 3 P002  Placebo    82 V0      6.6    26     0
#> 4 P002  Placebo    82 V3      8.5    27    36
#> 5 P003  Placebo    67 V0      4.9    28     0
#> 6 P003  Placebo    67 V3      5.2    27    36