Pre-processing data

“Data” preprocessing refers to the code which contains:

Data import calls
Data modification

Including the preprocessing code is an important step that is handled by teal.data functions.

For example, for the following example only ADSL <- readRDS("<your data path>/adsl.rds") is considered as preprocessing code:

library(scda)
library(teal.data)
saveRDS(synthetic_cdisc_data("latest")$adsl, "adsl.rds")

## preprocessing -------------------
adsl <- readRDS("adsl.rds")
## -------------------
data <- cdisc_data(cdisc_dataset("ADSL", adsl))
data$get_code()

## [1] ""

If you run the example above, the get_code function will return an empty string reflecting that “Preprocessing is empty”. In order to show the preprocessing code correctly the code argument of the cdisc_data function needs to be specified. For the above example this would be:

library(scda)
saveRDS(synthetic_cdisc_data("latest")$adsl, "adsl.rds")

## preprocessing -------------------
adsl <- readRDS("adsl.rds")
## -------------------
unlink("adsl.rds")


data <- cdisc_data(
  cdisc_dataset("ADSL", adsl),
  code = 'ADSL <- readRDS("adsl.rds")'
)
data$get_code()

## [1] "ADSL <- readRDS(\"adsl.rds\")"

The code used to get the ADSL dataset is returned as expected. This can be used as input to the data argument of teal::init to ensure reproducibility in teal apps.

NEST CoreDev

2022-07-26