R 4.0

[pcorvalan]

Hi Steven,

I would like to ask if you test Salesforcer on R 4.0 environment.
Does Salesforcer have any dplyr package dependency that may be affected if we upgrade to dplyr 1.0?

Saleforcer is a great package!

Thanks!

[StevenMMortimer]

@pcorvalan Great question. I haven't tested on R 4.0. The current tests were on minor versions R 3.5 and dplyr 0.8.

I wouldn't expect much to change on R 4.0. Most of the references to stringsAsFactors already involved setting it to FALSE.

I'm not so sure about dplyr 1.0. My advice is taking a look at the changes in the recent release to spot any potential issues, then upgrade in a dev environment see how it goes. If not, then revert and let me know any issues. Happy to help work through it to make sure it works for everyone.

[arp3ggio]

Fwiw, I just upgraded today to R 4.0 today on Ubuntu 20.04. I had a few battles to fight getting various tidyverse packages re-installed, but salesforcer was not a problem. I did receive the following message on my first run:
Error: package ‘salesforcer’ was installed before R 4.0.0: please re-install it

From terminal, I then ran:
sudo su - -c "R -e \"devtools::install_github('StevenMMortimer/salesforcer', force = TRUE)\""

The force = TRUE was required because (I believe) devtools::install_github() was detecting no changes since prior install. Prior to the above, I also had to re-install base64enc and data.table per ERROR messages on my initial attempts to reinstall salesforcer. I likely re-installed the other dependencies listed here as part of reinstalling other packages:
https://cran.r-project.org/web/packages/salesforcer/index.html

Overall, I could have been a bit more organized before doing the upgrade to R 4.0, but I did full disk clone last night, so I was feeling reckless.

In minimal initial testing (numerous sf_query but nothing else), I'm not seeing any issues.

[StevenMMortimer]

Thanks, @arp3ggio! Appreciate the detail here. Very helpful. Keep me posted if anything pops up.

[pcorvalan]

Thanks @arp3ggio !!

[arp3ggio]

Additional background, I'm also using dplyr version 0.8.99.9003 from github. Since my prior post, I've not seen any problems with numerous sf_create.

[danwwilson]

I was just trying to use sf_query() with the bulk API and got the following error

Error: Unknown TZ UTC
In addition: Warning message:
In OlsonNames() : no Olson database found

I did some tracing and the error occurs at line 52 of sf_create_job_bulk_v1() with the statement

job_info <- response_parsed %>% xml_ns_strip() %>% xml_find_all("//jobInfo") %>% 
    map_df(xml_nodeset_to_df) %>% type_convert(col_types = cols())

I had the issue myself the other day with another function of mine as the result of using a posixct in an interval function, but given this issue is the result of the type_convert it could be due to some sort of missing timezone information.

[StevenMMortimer]

@danwwilson After some googling, I think it could be more of a system configuration issue where readr cannot find the Olson database on the system needed to parse datetimes. Could you check out the two similar issues I've linked below and let me know if these explain/solve the issue? Thanks.

• Why does readr::read_csv need timezone data and Olson database? - Why does readr::read_csv need timezone data and Olson database? tidyverse/readr#952
• R missing timezone data and Olson database - R missing timezone data and Olson database jupyter/docker-stacks#502

[danwwilson]

@StevenMMortimer adding the tzdata to the docker build solved the problem. Thanks a Buch for your help.

[StevenMMortimer]

@pcorvalan @arp3ggio @danwwilson The latest release of the package is now on cran (salesforcer 0.1.4). While releasing I tested and built the package on Windows, Ubuntu, and Mac OS X across multiple version of R, namely 3.6.3, 4.0.0, and 4.0.1 (more specifics on the tested configurations are available in cran-comments.md). I didn't encounter any problems with unit tests or installs. As @arp3ggio keenly pointed out, you may have to work through some issues installing the dependencies, but no trouble after that.

Please let me know if there are still problems and we can work through them. If there are no lingering issues or I don't hear back in a couple weeks, then I'll close this issue. Other updates made in v0.1.4 are available in NEWS.md

Thanks again for your assistance and your patience! 👍

[pcorvalan]

Thanks Steven! We upgraded R base in a test machine without any problem.
Best regards.