This isn't a major issue as fread can be used within read_using, but fread is faster than read.csv and copes well with untidy CSVs (e.g. where the first row has fewer columns than the rest, which unfortunately occurs in downloads of SOP reports).
Both fread and read_csv are fast and apply similar default options to read_excel (e.g. trimming white space), but read_csv doesn't cope so well with untidy CSVs.
I realise either way it creates a dependency on another package, which may be better avoided.
This isn't a major issue as fread can be used within read_using, but fread is faster than read.csv and copes well with untidy CSVs (e.g. where the first row has fewer columns than the rest, which unfortunately occurs in downloads of SOP reports).
Both fread and read_csv are fast and apply similar default options to read_excel (e.g. trimming white space), but read_csv doesn't cope so well with untidy CSVs.
I realise either way it creates a dependency on another package, which may be better avoided.