How to save time zone information in R-CodePudding

I am trying to find out if it is possible to save the date time zone into a file in R. Lets create a date/time variable set to EST time.

x <- as.POSIXct("2015-06-01 12:00:00", tz = "EST")
x
#> [1] "2015-06-01 12:00:00 EST"

lubridate::tz(x)
#> [1] "EST"

tmpfile <- tempfile()

I will save this date information (which is in EST time) using base R, readr and data.table to see how the information is preserved.

Base R version

The date gets converted to UTC and the time is not changed.

write.csv(data.frame(x = x), tmpfile)
df <- read.csv(tmpfile)

df
#>   X                   x
#> 1 1 2015-06-01 12:00:00

lubridate::tz(df$x)
#> [1] "UTC"

tibble/readr version

It gets converted to UTC and the time is formatted as UTC as described in the help. Time is now 17:00:00.

… POSIXct values are formatted as ISO8601 with a UTC timezone Note: POSIXct objects in local or non-UTC timezones will be converted to UTC before writing.

readr::write_csv(tibble::tibble(x = x), tmpfile)
df <- readr::read_csv(tmpfile)
#> Rows: 1 Columns: 1
#> ── Column specification ────────────────────────────────────────────────────────
#> Delimiter: ","
#> dttm (1): x
#> 
#> ℹ Use `spec()` to retrieve the full column specification for this data.
#> ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

df
#> # A tibble: 1 × 1
#>   x                  
#>   <dttm>             
#> 1 2015-06-01 17:00:00

lubridate::tz(df$x)
#> [1] "UTC"

data.table version

The date gets converted to UTC and the time is formatted as UTC as described in the help. Time is now also 17:00:00.

data.table::fwrite(data.table::data.table(x = x), tmpfile)
df <- data.table::fread(tmpfile)

df
#>                      x
#> 1: 2015-06-01 17:00:00

lubridate::tz(df$x)
#> [1] "UTC"

Is it possible to preserve the time zone when saving a file? If the user is not paying attention, this can go silently and create issues.

^{Created on 2022-03-14 by the reprex package (v2.0.1)}

CodePudding user response：

Saving to an .RDS is often the best choice to preserve an object exactly as it’s represented in R. In this case, it preserves timezone without conversion:

saveRDS(data.frame(x = x), tmpfile)
df <- readRDS(tmpfile)

df
#                    x
# 1 2015-06-01 12:00:00

lubridate::tz(df$x)
# [1] "EST"

df$x
# [1] "2015-06-01 12:00:00 EST"

CodePudding user response：

You could simply paste the time zone on to the end of the datetime string.

x <- as.POSIXct("2015-06-01 12:00:00", tz = "EST")
df <- data.frame(x = paste(as.character(x), lubridate::tz(x)))

tmpfile <- tempfile()

write.csv(df, tmpfile, row.names = FALSE)

new_df <- read.csv(text = csv)

new_df
#> x
#> 1 2015-06-01 12:00:00 EST

as.POSIXct(new_df$x, strsplit(new_df$x, ' ')[[1]][3])
#> [1] "2015-06-01 12:00:00 EST"