[Feature Request]: Validate the data summary in the active filters

[vedhav]

Feature description

If a user provides incorrect join keys, they might make the filter logic produce incorrect data. One common pattern is when the parent-child mappings are reversed. This would make the child filter the parent and cause unusual behavior.

Example app to reproduce this error:

library(teal)

data <- within(teal_data(), {
  library(scda)
  library(scda.2022)
  library(dplyr)
  library(nestcolor)

  ADSL <- synthetic_cdisc_data("latest")$adsl

  adsl_labels <- teal.data::col_labels(ADSL, fill = FALSE)
  ADSL <- ADSL %>%
    mutate(
      TRTDURD = as.numeric(as.Date(TRTEDTM) - as.Date(TRTSDTM)) + 1,
      DTHFL = ifelse(!is.na(DTHDT), "Y", NA),
      EOSSTT = factor(EOSSTT, levels = c("COMPLETED", "ONGOING", "DISCONTINUED"))
    ) %>%
    teal.data::col_relabel(
      TRTDURD = "Treatment Duration in Days",
      DTHFL = "Death Flag",
      DCSREAS = "Reason for Study Discontinuation",
      EOSSTT = "End of Study Status"
    ) %>%
    droplevels()
  teal.data::col_labels(ADSL)[c(names(adsl_labels))] <- adsl_labels

  ADRSSWIM <- synthetic_cdisc_data("latest")$adrs %>%
    filter(PARAMCD == "OVRINV") %>%
    arrange(USUBJID)
})

datanames <- c("ADSL", "ADRSSWIM")
datanames(data) <- datanames

join_keys(data) <- default_cdisc_join_keys[datanames]
join_keys(data)["ADRSSWIM", "ADRSSWIM"] <- c("STUDYID", "USUBJID", "PARAMCD", "AVISIT")


# Broken data in filter panel
join_keys(data)["ADRSSWIM", "ADSL"] <- c("STUDYID", "USUBJID")

# Correct mapping to fix the issue
# join_keys(data)["ADSL", "ADRSSWIM"] <- c("STUDYID", "USUBJID")

app <- init(
  data = data,
  modules = example_module()
)

shinyApp(app$ui, app$server)

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

Security Policy

• I agree to follow this project's Security Policy.

[chlebowa]

I don't quite understand the error. The app seems to be working fine. Can you please elaborate?

[vedhav]

Added a screenshot in the description. The problem is with the improper joins in the filter panel. There are 2400 rows of ADSL created which are just duplicates of the 400 actual rows.
We should have an error or warning when when invalidations occur due to bad user input, in this case improper parent-child mapping.

[chlebowa]

Got it, thank you 👍