Failed to parse datetime with a space in format

[hongyuanjia]

Possibly related to #1276.

# development version
library(readr)
packageVersion("readr")
#> [1] '2.0.1.9000'

fmt_space <- "%d/%m/%y %H:%M"
fmt_nospace <- "%d/%m/%y%H:%M"

# both work using parse_datetime()
with_edition(1, readr::parse_datetime("1/9/19 0:00", fmt_space))
#> [1] "2019-09-01 UTC"
with_edition(2, readr::parse_datetime("1/9/19 0:00", fmt_space))
#> [1] "2019-09-01 UTC"
with_edition(1, readr::parse_datetime("1/9/19 0:00", fmt_nospace))
#> [1] "2019-09-01 UTC"
with_edition(2, readr::parse_datetime("1/9/19 0:00", fmt_nospace))
#> [1] "2019-09-01 UTC"

# create example data
s <- c("datetime, value", "1/9/19 0:00, 1")
f <- tempfile(fileext = "csv")
writeLines(s, f)

# both work with edition 1
with_edition(1,
  readr::read_csv(
    f,
    col_types = cols(
      col_datetime(fmt_space),
      col_integer()
    )
  )
)
#> # A tibble: 1 x 2
#>   datetime            value
#>   <dttm>              <int>
#> 1 2019-09-01 00:00:00     1
with_edition(1,
  readr::read_csv(
    f,
    col_types = cols(
      col_datetime(fmt_nospace),
      col_integer()
    )
  )
)
#> # A tibble: 1 x 2
#>   datetime            value
#>   <dttm>              <int>
#> 1 2019-09-01 00:00:00     1

# however, it fails in edition 2 if there is a space
with_edition(2,
  readr::read_csv(
    f,
    col_types = cols(
      col_datetime(fmt_space),
      col_integer()
    )
  )
)
#> Warning: One or more parsing issues, see `problems()` for details
#> # A tibble: 1 x 2
#>   datetime            value
#>   <dttm>              <int>
#> 1 NA                      1

# without spacing, it works
with_edition(2,
  readr::read_csv(
    f,
    col_types = cols(
      col_datetime(fmt_nospace),
      col_integer()
    )
  )
)
#> # A tibble: 1 x 2
#>   datetime            value
#>   <dttm>              <int>
#> 1 2019-09-01 00:00:00     1

^{Created on 2021-09-06 by the reprex package (v2.0.0)}

[jimhester]

This should be fixed by installing the latest version of vroom (1.5.5)

[hongyuanjia]

Thanks! Just checked and the problem goes away after updating vroom to v1.5.5.