Open LukasWallrich opened 3 years ago
Hi there!
Thanks so much for posting a reprex for this issue.
It was indeed @andrewheiss who implemented this, so thanks to him! :)
I confirm that I am getting this issue as well,
gapminder <- tibble::tribble(
~life_expectancy, ~fertility, ~population, ~gdp,
72.7, 1.9, 360830, 6650151639,
NA, 2.14, 1261319, 14178489076,
68.8, 2.28, 151616777, 82979485251,
75.2, NA, 279566, NA,
70.2, 1.46, 9492122, 26002131603,
80.1, 1.84, 10929978, 267143096398,
71.2, NA, 321609, 1221526439,
60.8, 5.1, 9509798, 3336801340,
71.4, 2.38, 720246, 961365501.9,
71.8, 3.36, 9918245, 12240412986
)
naniar::mcar_test(gapminder)
#> Error: Problem with `mutate()` input `d2`.
#> x system is computationally singular: reciprocal condition number = 5.00265e-23
#> ℹ Input `d2` is `purrr::pmap_dbl(...)`.
#> ℹ The error occurred in group 1: miss_pattern = 1.
Created on 2021-04-20 by the reprex package (v2.0.0)
I've had a bit of a poke around, and it looks like this is happening when running solve
here.
This sort of error can happen when you cannot invert a matrix - see this SE thread.
@andrewheiss have you run into this issue?
When I have time, I'll take a look at providing some suggested workarounds for this kind of problem, and trying to document some ways to avoid this.
I have run into this issue, but haven't figured out a way around it yet. ☹️
Hello everyone, everything good? Has anyone found a solution to this problem? I have yet to see any so far.
@andrewheiss Has anyone looked into this? I keep receiving the following error messages for just one group of my dataset:
Error in `dplyr::mutate()`:
ℹ In argument: `d2 = purrr::pmap_dbl(...)`.
ℹ In group 1: `miss_pattern = 1`.
Caused by error in `purrr::pmap_dbl()`:
ℹ In index: 1.
Caused by error in `solve.default()`:
! Lapack routine dgesv: system is exactly singular: U[30,30] = 0
Thanks for adding
mcar_test()
. Unfortunately, the function fails for me due to singularity, regardless of how I slice my data. I am not sure whether that is a limitation of the test (which might then benefit from a clearer error message or documentation) or an issue with the implementation? Below a reprex with a small subset of the data, but I get the same issue with 185 rowsCreated on 2021-04-07 by the reprex package (v1.0.0)