results_get() doesn't work

[leeper]

The results_get() function doesn't work.

[leeper]

This still doesn't work. While it's possible to specify type = "aggregated", type = "full" produces an error due to some of the internal parsing performed by rowDataToDF(). @pablobarbera maybe you can figure out what's going wrong here?

[pablobarbera]

I added a bunch of changes to rowDataToDF() that should fix all these problems. The commit has a few more details on what the issues were -- mostly edge cases and dealing with multiple choice answers. Very ugly code, but it should work (or at least it did with all the different jobs I have). If you can try with your jobs as well, to make sure I didn't miss anything, that would be great. :)

[leeper]

I'm still getting the same kind of error:

> results_get(859922, type = "full")
2 rows downloaded
Error in `[<-.data.frame`(`*tmp*`, , var, value = 1890849317L) : 
  replacement has 1 row, data has 0

I'll try to debug.

[pablobarbera]

Ugh. Could you perhaps share the output of results_get() before it calls rowDataToDF()? It must be another edge case with a specific question design or something like that. I can look into this.

[vanatteveldt]

I encountered this problem as well. In my case, the job had a number of variables that were almost never used, which resulted in columns not existed. The existing check in rowDataToDF checked for NA, but not for NULL. Adding a NULL check seemed to solve the problem for me (see pull request above)