db_append_table should not double escape already escaped quotes

[jd4ds]

Hey there,
long awaited and now available, I was very happy about the possibility to insert data into MySQL databases via local file. However, I may have found a small issue when creating the TSV file.
{readr} seems to escape already escaped quotes, which leads to unwanted results. My suggestion would be to use readr::write_delim(..., escape = "none"), unless the default escaping of readr was chosen on purpose.

Please let me know if you need more information on this!

Cheers,
Janis

escape_testing <- function(escape = "double"){
  file_default_escape <- tempfile(fileext = ".tsv")
  readr::write_delim(
    RMariaDB:::csv_quote(value,
                         warn_factor = TRUE,
                         conn = target_conn),
    file_default_escape,
    quote = "none",
    delim = "\t",
    na = "\\N",
    col_names = FALSE,
    escape = escape
  )
  value_default_escape <- readr::read_tsv(file = file_default_escape)
  print(as.data.frame(value_default_escape))
  unlink(file_default_escape)
}

value <-
  data.frame(test = c("test_header", 'Test \"quote\" escaping.'))

# Escape default value
escape_testing(escape = "double")
#> Rows: 1 Columns: 1
#> -- Column specification --------------------------------------------------------
#> Delimiter: "\t"
#> chr (1): test_header
#> 
#> 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.
#>                test_header
#> 1 Test ""quote"" escaping.

# `escape = "none"`
escape_testing(escape = "none")
#> Rows: 1 Columns: 1
#> -- Column specification --------------------------------------------------------
#> Delimiter: "\t"
#> chr (1): test_header
#> 
#> 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.
#>              test_header
#> 1 Test "quote" escaping.

[jd4ds]

To add on this a little:
readr::write_delim still seems way slower than for instance to use data.table::fwrite. Since this function is suppose to focus on speed, would it be an option to consider swapping the function that creates the tsv?

[krlmlr]

Thanks, happy to review a PR, I don't have a strong opinion regarding the method to create the TSVs as long as it's robust.

[jd4ds]

Sure I can look into that.
But for now is there any reason why to use the default escaping from {readr}? Otherwise I would suggest just to add the escape = "none" argument to the function call as a quick fix (for now). ( I could open a PR for this as well if requested).
Thanks!

[jd4ds]

I tested the changes against all our MySQL database schemas for identity and it seems to work just fine. I do not really know how DBItest works and on a first glance it seems the issue is with MariaDB which I do not have access to.
Anything I can do here?