is.file.binary {knitrdata}

Functions to assess if files are binary, DOS text or UNIX text format

Description

These functions attempt to determine if a file is binary or text. In addition, file.type attempts to determine the newline character(s) used in the file.

Usage

is.file.binary(file, bin.ints = c(1:8, 14:25), nbytes = 1000, nbin = 2)

file.type(file, bin.ints = c(1:8, 14:25), nbytes = 1000, nbin = 2)

Arguments

file

The path to the file to be examined

bin.ints

List of integers with the ASCII values of control characters that are to be considered when when looking for signs a file is binary. Default includes most ASCII control characters except things like NULL, LF, CR and HT that might actually appear in an ASCII file.

nbytes

Number of bytes to read in from the beginning of the file.

nbin

An integer indicating the threshold on the number of control characters above which a file is considered binary. Defaults to 2.

Details

A file is assessed to be binary using a heuristic based on finding more than nbin ASCII control (i.e., non-printing) characters in the first nbytes of the file. This works well for standard ASCII text, but it may be less effective for complex UTF8 text (e.g., Chinese).

For text files, line endings are assessed by file.type by searching first for DOS line endings (\r\n) in the first nbytes of the input file, and then by searching for UNIX line endings (\n). If neither is found, then NA_character_ is returned for the line ending.

Value

For is.file.binary, a boolean value is returned, whereas a list is returned for file.type.

Functions

• is.file.binary: A boolean that will be TRUE if a file is considered to be binary.

• file.type: Returns a list with up to two elements: type & newline. type can either by "binary" or "text". newline will be NULL for binary files, "\r\n" for DOS formatted text files, "\n" for UNIX formatted text files and NA_character_ for text files without any newline characters in the first nbytes of the file.

Author(s)

David M. Kaplan [e-mail]

See Also

See also platform.newline.