lenses and stype

The stype package provides lenses for getting and settings two types of data:

  1. the attributes of an individual stype vector (e.g. a context, purpose or internal_name)
  2. sets of stype vectors within a list-like structure (such as a data.frame or tibble)

view and set elements of a stype vector


A lens consists of two functions: view and set. set can be used to update parts of a stype vector.

x <-  v_binary() %>%
  set(derivation_l, "some derivation") %>%
  set(long_label_l, "a new label") %>%
  set(study_role_l, "outcome")
#> <binary[0]>
#> Proportion = NaN
#> Purpose: outcome

And the view can be used to get parts of a vector:

view(x, derivation_l)
#> [1] "some derivation"
view(x, long_label_l)
#> [1] "a new label"
view(x, study_role_l)
#> [1] "outcome"

In most cases, stype also has get_* functions that mimic view:

#> [1] "some derivation"
#> [1] "a new label"

view and set stype vectors within a list-like structure

Lenses for list-like structures can be used to view and set vectors within a data.frame. In this example, we set all continuous variables to be covariates, binary to exposures, and count to outcome:

df <- tibble(
  x1 = v_binary(),
  x2 = v_continuous(),
  x3 = v_continuous(),
  x4 = v_count()

df <- df %>%
  over_map(continuous_l, function(x) set(x, study_role_l, "covariate"))%>%
  over_map(binary_l, function(x) set(x, study_role_l, "exposure")) %>%
  over_map(count_l, function(x) set(x, study_role_l, "outcome"))

view(df$x1, study_role_l)
#> [1] "exposure"
view(df$x2, study_role_l)
#> [1] "covariate"
view(df$x3, study_role_l)
#> [1] "covariate"
view(df$x4, study_role_l)
#> [1] "outcome"