In EdwinTh/checkr:

This is a thinking through how the package should work, before doing any implementations. A first idea is to implement many individual units with which the user can make the assumptions to the data, both within a function as input and output checking, and in individual scripts. For the inputs and the outputs there should be wrapper functions that checks all the arguments. check_inputs will check the input variables and will return norhing if passes. check_outputs does the same, but does return the input when passes.

Some sketches:

Unit functions:

has_df_class(mtcars)
is_char_type("a")
has_fctr_class()
has_fctr_levels
has_length()
has_nrow(mtcars, 10)
has_df_schema()
has_names()

Within a function:

check_inputs({
  x %>% has_df_class() %>%
    has_colnames("id", "date", "score") %>% 
    has_df_schema(id = integer, date = Date, score = double)
  y %>% is_list_type() %>% has_names("analysis1", "analysis2")
})

check_outputs(
  x %>% is_numeric_type() %>% has_length(100)
)

Helper functions:

get_df_schema(mtcars)

EdwinTh/checkr documentation built on March 29, 2020, 6:37 a.m.