Plain text formats (2)

The functions work faster if prespecifying column types.

If not prespecifying column types, readr and data.table make a guess (based on first 1000 rows).

For columns with first 1000 rows as NA:

• at the time of writing the book, such column was set as logical by readr::read_csv()
• nowadays such column is assumed to be character

There’s also read_csv_arrow(), read_tsv_arrow(), read_delim_arrow() and read_json_arrow() from the arrow package: see further.