sortbib - sort bibliographic database
sortbib [ -sKEYS ] database ...
Sortbib sorts files of records containing refer key-letters by user-specified keys.
Records may be separated by blank lines, or by .[ and .] delimiters, but the two styles
may not be mixed together. This program reads through each database and pulls out key
fields, which are sorted separately. The sorted key fields contain the file pointer, byte
offset, and length of corresponding records. These records are delivered using disk seeks
and reads, so sortbib may not be used in a pipeline to read standard input.
By default, sortbib alphabetizes by the first %A and the %D fields, which contain the
senior author and date. The -s option is used to specify new KEYS. For instance, -sATD
will sort by author, title, and date, while -sA+D will sort by all authors, and date.
Sort keys past the fourth are not meaningful. No more than 16 databases may be sorted
together at one time. Records longer than 4096 characters will be truncated.
Sortbib sorts on the last word on the %A line, which is assumed to be the author's last
name. A word in the final position, such as ``jr.'' or ``ed.'', will be ignored if the
name beforehand ends with a comma. Authors with two-word last names or unusual construc-
tions can be sorted correctly by using the nroff convention ``\0'' in place of a blank. A
%Q field is considered to be the same as %A, except sorting begins with the first, not the
last, word. Sortbib sorts on the last word of the %D line, usually the year. It also
ignores leading articles (like ``A'' or ``The'') when sorting by titles in the %T or %J
fields; it will ignore articles of any modern European language. If a sort-significant
field is absent from a record, sortbib places that record before other records containing
refer(1), addbib(1), roffbib(1), indxbib(1), lookbib(1)
Greg Shenaut, Bill Tuthill
Records with missing author fields should probably be sorted by title.
4.2 Berkeley Distribution April 29, 1985 SORTBIB(1)