Closed ablack3 closed 1 week ago
Also had an error when running this on the ohdsi test postgres database.
@ablack3 In this context covariates
is an Andromeda object so I'm wondering if this is related to any recent changes there and if we need to make alterations to support them? I had problems with the shelved Arrow backend usage, are there any changes to the interface with regard to this?
No changes to Andromeda but I think dbplyr had a release recently so maybe that is the cause. I'll try to dig into it and post my findings.
Well I have not had much time to dig into why this is failing (lots to do 🚀 ) but I did download the current release and run the devtools::test()
on sqlite only. There were a couple test failures but a large number of warnings so I was surprised that this is passing R check. But actually when I check the Hades website I see it is failing R check according to the dashboard.
I know I have some strong opinions, maybe too many (and I absolutely don't mean to be overly critical and am so appreciative of the Hades codebase and maintainers and I'm as bad as anyone about not passing tests), but I really think that putting R packages on CRAN is a good idea because it prevents this situation. If you don't pass R check (which includes your tests) you get kicked off after a short period of time. So if you want to be on CRAN you have to pass your own tests. It's a good thing to enforce I think. In Hades you can fail R check and remain as the current release which I don't think is good practice.
Backtrace:
▆
1. ├─CohortDiagnostics::uploadResults(...) at test-1-ResultsDataModel.R:163:4
2. │ └─ResultModelManager::uploadResults(...) at CohortDiagnostics/R/ResultsDataModel.R:135:2
3. │ └─base::lapply(unique(specifications$tableName), uploadTable)
4. │ └─ResultModelManager (local) FUN(X[[i]], ...)
5. │ └─readr::read_csv_chunked(...)
6. │ └─readr:::read_delimited_chunked(...)
7. │ └─readr:::read_tokens_chunked(...)
8. │ └─readr:::read_tokens_chunked_(...)
9. └─readr (local) `<fn>`(`<spc_tbl_[,7]>`, 1L)
10. └─private$callback(data, index)
[ FAIL 2 | WARN 23 | SKIP 0 | PASS 103 ]
@ablack3 when using the latest version of OHDSI CohortDiagnostics, I am not getting the error from your reprex anymore. This seems to be the fix
Appears to be fixed
I'm not sure if I'm doing something wrong here or what this error is about but it seems like my small example code should work.
Created on 2023-09-01 with reprex v2.0.2