Undefined columns selected error

[batpigandme]

As discussed in #2584, this is an issue for seven revdeps.

Reprex from ggseas package

library(ggseas)
#> Loading required package: seasonal
#> Loading required package: ggplot2
#> Loading required package: zoo
#> 
#> Attaching package: 'zoo'
#> The following objects are masked from 'package:base':
#> 
#>     as.Date, as.Date.numeric
ap_df <- tsdf(AirPassengers)
ggsdc(ap_df, aes(x = x, y = y), method = "decompose") +
  geom_line()
#> Error in `[.data.frame`(data, , xvar): undefined columns selected

Traceback

8: stop("undefined columns selected")
7: `[.data.frame`(data, , xvar)
6: data[, xvar]
5: order(data[, xvar])
4: `[.data.frame`(data, order(data[, xvar]), )
3: data[order(data[, xvar]), ]
2: ggsdc_helper(data = data, mapping = mapping, method = method, 
       s.window = s.window, type = type, index.ref = index.ref, 
    ...
1: ggsdc(ap_df, aes(x = x, y = y), method = "decompose")

Same basic structure for scatterpie:

ggplot() + geom_scatterpie(aes(x=x, y=y), data=d, cols=c("A", "B", "C")) + coord_fixed()
#>    Error in `[.data.frame`(data, , xvar) : undefined columns selected
#>    Calls: geom_scatterpie -> diff -> [ -> [.data.frame
#>    Execution halted

[hadley]

Scatterpie has the same root cause: https://github.com/GuangchuangYu/scatterpie/blob/master/R/geom_scatterpie.R#L33

So I suspect the problem is that some extension authors aren't clear about the correct way to accomplish their goals, and we'll need to help them.

[batpigandme]

Looks like that's the same as what's in gggenes, too

dummies <- make_alignment_dummies(example_genes, ggplot2::aes(xmin = start,
    +   xmax = end, y = molecule, id = gene), on = "genE")
#>    Error in `[.data.frame`(data, as.character(unlist(mapping))) : 
#>      undefined columns selected
#>    Calls: make_alignment_dummies -> [ -> [.data.frame
#>    Execution halted

[hadley]

Here that involves using rlang::quo_name() instead of as.character() (which I'm pretty sure wouldn't have work before with expressions like x + y). And in general (except in very special circumstances), you shouldn't be looking for aesthetics in the data — typically you should be implementing a layer function where the aesthetics have already been mapped.

[cperk]

Is there any way to resolve this error when it occurs with scatterpie?

[thomasp85]

You’ll need to reach out to the developer of scatterpie.

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/