Invalid timezone of the input vector: "GMT" using Linear relations between pollutants function

[cross85]

I was checking the linear relation between pollutants function with the test data and I've got this error message:

Error: Problem with `mutate()` column `cond`.
i `cond = round_date(date, period)`.
x CCTZ: Invalid timezone of the input vector: "GMT"

After that I run rlang::last_error() and I get this message:

<error/dplyr:::mutate_error>
Problem with `mutate()` column `cond`.
i `cond = round_date(date, period)`.
x CCTZ: Invalid timezone of the input vector: "GMT"

Then I checked the backtrace rlang::last_trace()

<error/dplyr:::mutate_error>
Problem with `mutate()` column `cond`.
i `cond = round_date(date, period)`.
x CCTZ: Invalid timezone of the input vector: "GMT"
Backtrace:
     x
  1. +-openair::linearRelation(mydata, x = "nox", y = "so2")
  2. | +-mutate(mydata, cond = round_date(date, period)) %>% group_by(cond)
  3. | +-dplyr::mutate(mydata, cond = round_date(date, period))
  4. | \-dplyr:::mutate.data.frame(mydata, cond = round_date(date, period))
  5. |   \-dplyr:::mutate_cols(.data, ..., caller_env = caller_env())
  6. |     +-base::withCallingHandlers(...)
  7. |     \-mask$eval_all_mutate(quo)
  8. +-dplyr::group_by(., cond)
  9. +-lubridate::round_date(date, period)
 10. | +-base::as.POSIXct(ceiling_date(x, unit = unit, week_start = week_start))
 11. | \-lubridate::ceiling_date(x, unit = unit, week_start = week_start)
 12. |   +-stats::update(x, seconds = second(x) - 1e-05, simple = T)
 13. |   \-lubridate:::update.POSIXt(x, seconds = second(x) - 1e-05, simple = T)
 14. |     +-base::do.call(...)
 15. |     \-(function (object, years = integer(), months = integer(), days = integer(), ...
 16. |       +-lubridate::reclass_date(do.call(C_update_dt, updates), object)
 17. |       +-lubridate:::reclass_date.POSIXct(...)
 18. |       | \-base::as.POSIXct(new)
 19. |       +-base::do.call(C_update_dt, updates)
 20. |       \-(function (dt, year, month, yday, mday, wday, hour, minute, second, ...
 21. +-base::stop(...)
 22. \-(function (e) ...
<error/Rcpp::exception>
CCTZ: Invalid timezone of the input vector: "GMT"

I'm using R 4.1.0 and openair 2.8-3
I tried using my own data and I've got the same error message
I just updated to the latest versions of R and openair, is this some incompatibility or some bug on the openair script?

[davidcarslaw]

Strange. Works for me OK. Have you updated all other packages through update.packages()? I get:

openair::linearRelation(mydata, x = "nox", y = "so2")

[cross85]

I did update.packages() but I think nothing related to openair was updated.
The problem is when "GMT" is set. As example.

If I import the data this wasy, I have no problem:

PM10 <- read_delim("PM10.csv", ";", escape_double = FALSE, 
col_types = cols(date = col_datetime(format = "%d/%m/%Y %H:%M")), trim_ws = TRUE)

But If I import the data with no format, and I apply it after with GMT as time zone, I get the error message.

PM10 <- read_delim("PM10.csv", ";", escape_double = FALSE, trim_ws = TRUE)
PM10$date <- as.POSIXct(strptime(PM10$date, format = "%d/%m/%Y %H:%M", "GMT"))

[davidcarslaw]

mmm ... yes. I think this is something to do with how lubridate handles base R times. Does this work:

library(lubridate)
PM10$date <- dmy_hm(PM10$date, tz = "GMT"))

[cross85]

I think maybe it's lubridate problem. I found this issue checking on lubridate, it might be that library issue, apparently with the folder where the time zones are stored. #928

I tried what you said and I've got the same error messages. So, for now I will use it without setting a time zone.

> library(readr)
> PM10 <- read_delim("PM10.csv", 
+     ";", escape_double = FALSE, trim_ws = TRUE)
> library(lubridate)

Attaching package: ‘lubridate’

The following objects are masked from ‘package:base’:

    date, intersect, setdiff, union

> PM10$date <- dmy_hm(PM10$date, tz = "GMT")
> str(PM10)
spec_tbl_df [8,712 x 10] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
 $ date   : POSIXct[1:8712], format: "2019-01-01 01:00:00" "2019-01-01 02:00:00" "2019-01-01 03:00:00" "2019-01-01 04:00:00"
> library(openair)
> openair::linearRelation(PM10, x = "PM10_LC", y = "PM10_CH")
Error: Problem with `mutate()` column `cond`.
i `cond = round_date(date, period)`.
x CCTZ: Invalid timezone of the input vector: "GMT"
Run `rlang::last_error()` to see where the error occurred.

I'm using this version of R and openair 2.8-3 in windows with RStudio and lubridate 1.7.10

R version 4.1.0 (2021-05-18) -- "Camp Pontanezen"
Copyright (C) 2021 The R Foundation for Statistical Computing
Platform: x86_64-w64-mingw32/x64 (64-bit)

[jack-davison]

Closing as this appeared to be an issue with the {lubridate} package; if your issue is persisting do feel free to reopen and we'll take another look. Cheers!