as.data.frame(ds, force = TRUE) returns many columns of NA

[rossellhayes]

Starting with the recent update to the crunch package, running as.data.frame(ds, force = TRUE) now returns a data frame with many (sometimes 100s or even 1000s of) columns with NA names.

It appears that each of these NA columns come after an array variable. Where they occur, the array variable column has a length of 1, instead of the expected number of rows. Each of the NA columns also has length 1, and there is 1 NA column for each row of the dataset (minus the one named column). As a result, it seems that certain columns are ending up in the data horizontally rather than vertically, like so:

 a  b  c NA NA NA  d
 1  1  1  2  3  4  1
 2  2              2
 3  3              3
 4  4              4

This problem may be related to the recent addition of the array_strategy argument, but it seems to happen regardless of what array_strategy is set to.

[gergness]

Sorry, I believe this only happened when there was an array variable with a single subvariable, and then it was put horizontally (describing why there were so many columns).

Can you make sure that remotes::install_github("#663") remotes::install_github("Crunch-io/rcrunch#664") fixes for you too?

[rossellhayes]

Yes, that PR fixes the issue. Thanks!