Issue with time zones being misapplied / date-times being off in fetch_survey()

[jmobrien]

Note to users: update #263 on Jun 20 introduced a bug that could lead to date/time variables in the downloaded response dataframe to be off by 1 or more hours. Times displayed would reflect UTC time wrongly interpreted in the local time zone, or some other zone if time_zone was given explicitly.

This bug did not arrive on CRAN, so only downloads made the development version between then and now would have their data affected. The issue should now be fixed.

Re-downloading may be the easiest fix. If that is not possible, you should be able to do the following to correct your response data frame (here called dat):

dat |>
  mutate(
    across(
       c(StartDate, EndDate, RecordedDate),
       ~lubridate::force_tz(.x, "UTC") |> 
           lubridate::with_tz(Sys.timezone()) # Replace Sys.timezone() with alternative if used during download
    )
  )

[juliasilge]

Always good when bugs get caught before a release! 🐛

[jmobrien]

Indeed. Speaking of which, do I need to get up to speed on any plans for the next release, in terms of update goals, timelines, or both?

[juliasilge]

I'll be honest @jmobrien that I don't have many grand plans or goals for a next release here. 🙈 I would guess not super soon, since there are not that a ton of user-facing changes since the last release yet and none of extremely urgent? That bug fix would be nice to get to folks before the end of the year, for sure, and by then there will probably be a few more updates.

[jmobrien]

That makes sense; I think I was feeling similarly. One goal I've been wondering about is whether one key goal before the next (substantial) release should be to clear out all the V2 stuff (column_map(), metadata() survey_questions()), updating where appropriate. That would help clarify the package and improve ongoing maintenance a lot, I'd think.

[juliasilge]

That would be amazing to get done!