timelyportfolio/sunburstR

No tidyselect variables were registered

andirey opened this issue · 6 comments

andirey commented

I used "sunburstR" package.
But now it starts generate error

rlang::last_error()

message: No tidyselect variables were registered
class: rlang_error
backtrace:

  1. shiny::runApp()
  2. shiny:::serviceApp()
  3. shiny:::flushReact()
  4. .getReactiveEnvironment()$flush()
  5. ctx$executeFlushCallbacks()
  6. base::lapply(...)
  7. shiny:::FUN(X[[i]], ...)
  8. shiny:::flushCallback()
  9. shiny:::run()
  10. ctx$run(.func)
  11. env$runWith(self, func)
  12. shiny:::contextFunc()
  13. shiny:::<observer:output$plot_races_sunburst>(...)
  14. shiny:::func()
  15. shiny:::orig(name = name, shinysession = self)
  16. shiny:::output$plot_races_sunburst(...)
  17. shiny:::origRenderFunc(...)
  18. func()
  19. sunburstR::sunburst(df, count = TRUE, legend = FALSE) app.R:562:4
  20. sunburstR:::csv_to_hier(data)
  21. d3r::d3_nest(df, value_cols = "size")
  22. dplyr::one_of(c(nonnest_cols[length(nonnest_cols)], value_cols))
  23. tidyselect::peek_vars()
  24. vars_env$selected %||% abort("No tidyselect variables were registered")
    Call rlang::last_trace() to see the full backtrace
    ==
    version of sunburstR is 2.1.2

How to fix it?

timelyportfolio commented

@andirey not sure but my first guess to fix would be to install the newest d3r with install.packages("d3r"). Please let me know if this fixes.

  • 👍3
andirey commented

Yes. After "d3r" it works. Thanks
Why do we need d3r btw?

timelyportfolio commented

@andirey d3r does the conversion into expected d3 network format https://github.com/timelyportfolio/sunburstR/blob/master/R/converters.R. This is more reliable than the original JavaScript based converter.

  • 👍3
andirey commented

OK. Thanks for the advice!
I hope it can be useful for someone else as well.

timelyportfolio commented

@andirey thanks for filing the issue. The newest tidyr required modifications because of api change. I modified d3r to accept both old and new api format. Also, I have found the data.frame based network format as illustrated in https://bl.ocks.org/timelyportfolio/e135afab3838dffd82226655245d3feb to be more workable than the original delimited structure.

  • 👍2
timelyportfolio commented

Going to close. Feel free to reopen.

  • 👍1