version 2.0.0: na argument of read_tsv() does allow empty cells to be parsed with no missing values/NAs

[pfreese]

The na argument of readr::read_tsv() and related functions does not properly allow for a cell between two consecutive tabs ("\t\t") to be parsed as "" instead of NA.

df_w_empty_cell <- data.frame(name = c("ben", ""),
                              age = c(5, 10),
                              stringsAsFactors = FALSE)
# .tsv has contents:
# name  age
# ben   5
#       10
tsv_file <- "/tmp/df_w_empty_cell.tsv"
readr::write_tsv(df_w_empty_cell, tsv_file)

# > df_loaded
# A tibble: 2 x 2
# name    age
# <chr> <dbl>
# ben       5
# NA       10     <- This NA should be "".
df_loaded <- readr::read_tsv(tsv_file,
                             na = "NA")

[jimhester]

Thank you for opening the issue with a reproducible example!

This was a bug in read_tsv() with readr 2.0.0, it was not correctly passing the na argument down.