d3_nest() fails with dev tidyr

[hadley]

library(d3r)

titanic_df <- Titanic %>% 
  data.frame() %>% 
  select(Class,Age,Survived,Sex,Freq)

d3_nest(titanic_df, value_cols="Freq", root="titanic")
#> Error: No tidyselect variables were registered

Would you mind taking a look? It's probably because I've accidentally broken some aspect of nest(), but I can't quite figure out what the code is supposed to be doing

[timelyportfolio]

@hadley I fixed with 4b276b3. I assume I will wait for newest tidyr to hit CRAN and then submit this new version of d3r to CRAN.

[hadley]

What was the root cause of the problem?

[timelyportfolio]

I ended up rewriting to work with new version, but I believe I could have fixed by changing the named argument data = to .data =.

[jennybc]

d3r section of tidyr revdep check:

https://github.com/tidyverse/tidyr/blob/master/revdep/problems.md#d3r

Even when I check against the dev version of d3r, I see a problem. That is, 4b276b3 does not seem to actually resolve things. I still see this:

══ testthat results  ═══════════════════════════════════════════════════════
[ OK: 14 | SKIPPED: 4 | WARNINGS: 0 | FAILED: 2 ]
1. Error: d3_nest works as json (@test_hier.R#17) 
2. Error: d3_nest works with multiple NA levels (@test_hier.R#69) 

I've made another PR that does. Good news: small change! Bad news: requires dev tidyr. I lay out the options for next steps in the PR.

[jennybc]

Fixing this should also fix echarts4r and sunburstR:

https://github.com/tidyverse/tidyr/blob/master/revdep/problems.md#echarts4r

https://github.com/tidyverse/tidyr/blob/master/revdep/problems.md#sunburstr

[timelyportfolio]

@hadley @jennybc this should now be fixed to work with both old and new tidyr syntax and has been released on CRAN. Thanks for all your help.

I know sunburstR is definitely solved with this change and fairly confident echarts4r will be fixed as well.

Please reopen if not solved.