[ARROW-17429] [R] Error messages are not helpful of read_csv_arrow with col_types option

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Component/s: R
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/32697
• Assignee: SHIMA Tatsuya
• Reporter: SHIMA Tatsuya
• Created: 16/Aug/22 09:15
• Updated: 11/Jan/23 11:50
• Resolved: 01/Oct/22 09:43

Description

The error message displayed when a non-convertible type is specified does not seem to help in the development version.

tbl <- tibble::tibble(time = c("1970-01-01T12:00:00+12:00"))
csv_file <- tempfile()
on.exit(unlink(csv_file))
write.csv(tbl, csv_file, row.names = FALSE)

arrow::read_csv_arrow(csv_file, col_types = "?", col_names = "x", skip = 1)
#> # A tibble: 1 × 1
#>   x
#>   <dttm>
#> 1 1970-01-01 00:00:00
arrow::read_csv_arrow(csv_file, col_types = "c", col_names = "x", skip = 1)
#> # A tibble: 1 × 1
#>   x
#>   <chr>
#> 1 1970-01-01T12:00:00+12:00
arrow::read_csv_arrow(csv_file, col_types = "i", col_names = "x", skip = 1)
#> Error in as.data.frame(tab): object 'tab' not found
arrow::read_csv_arrow(csv_file, col_types = "T", col_names = "x", skip = 1)
#> Error in as.data.frame(tab): object 'tab' not found

In arrow 9.0.0

tbl <- tibble::tibble(time = c("1970-01-01T12:00:00+12:00"))
csv_file <- tempfile()
on.exit(unlink(csv_file))
write.csv(tbl, csv_file, row.names = FALSE)

arrow::read_csv_arrow(csv_file, col_types = "?", col_names = "x", skip = 1)
#> # A tibble: 1 × 1
#>   x
#>   <dttm>
#> 1 1970-01-01 00:00:00
arrow::read_csv_arrow(csv_file, col_types = "c", col_names = "x", skip = 1)
#> # A tibble: 1 × 1
#>   x
#>   <chr>
#> 1 1970-01-01T12:00:00+12:00
arrow::read_csv_arrow(csv_file, col_types = "i", col_names = "x", skip = 1)
#> Error:
#> ! Invalid: In CSV column #0: CSV conversion error to int32: invalid value '1970-01-01T12:00:00+12:00'
arrow::read_csv_arrow(csv_file, col_types = "T", col_names = "x", skip = 1)
#> Error:
#> ! Invalid: In CSV column #0: CSV conversion error to timestamp[ns]: expected no zone offset in '1970-01-01T12:00:00+12:00'

Issue Links

• Blocked: ARROW-15602 [R][Docs] Update docs to explain how to read timestamp with timezone columns
• is fixed by: ARROW-17355 [R] Refactor the handle_* utility functions for a better dev experience
• is related to: ARROW-15599 [R] Convert a column as a sub-second timestamp from CSV file with the `T` col type option
• links to: GitHub Pull Request #13922