Column-oriented FITS output consists of a FITS file containing two HDUs (Header+Data Units); a primary one (required by the FITS standard) and a single extension of type BINTABLE containing the table data. Unlike normal FITS format however, this table consists of a single row in which each cell holds the data for an entire column.
This can be a more efficient format to work with when dealing with very large, and especially very wide, tables. The benefits are greatest when the file size exceeds the amount of available physical memory and operations are required which scan through the table using only a few of the columns (many common operations, for instance plotting two columns against each other, fall into this category). The overhead for reading and writing this format is somewhat higher than for normal FITS however, and other applications may not be able to work with it (though it is a legal FITS file), so in most cases normal FITS is a more suitable choice.
There are two variants of this format:
The handler behaviour may be modified by specifying
one or more comma-separated name=value configuration options
in parentheses after the handler name, e.g.
The following options are available:
date = true|false
Multiple tables may be written to a single output file using this format.
If no output format is explicitly chosen,
writing to a filename with
the extension "
.colfits" (case insensitive)
colfits-plus format for output.
The handler classes for these formats are
For row-oriented FITS output, see Section 3.7.1.