$Id: README,v 1.1 1998/05/29 20:45:14 carlos Exp $

This directory contains experimental re-implementations of some of
the original NoSQL operators, that have been re-witten in C and AWK
for the sake of speed. They are not yet documented. Please refer to
their '-help' command line option for more info on each of them.

Disclaimer: I'm awful at C programming. If anyone wants to improve
my C code I'll be pleased :-). Furthermore, these operators have
been compiled and tested *only* on Debian/GNU Linux 1.3 (kernel 2.0.x).

These faster operators offer a less user-friendly command line
sintax and possibly different command line options than the Perl
ones, but they are meant to be used from inside other programs,
like WWW CGI scripts, where speed does matter and user-friendliness
doesn't that much.

If you want to try them out, simply edit the Makefile and change
INSTDIR to point to the directory you want to install the programs
into, then run:

                     make ; make install

from within this directory.

Note: some of these operators, namely nsq-fjoin and nsq-look,
need that the input table be split into two files, one with
the table header and another one with the table body. This can
be accomplished very easily with the commands :

    nsq-headchg -del < rdbtable > rdbtable.Body
    nsq-headchg -copy rdbtable < /dev/null > rdbtable.Header

Thanks to the great flexibility of the Operator/Stream paradigm,
a table thus split can still be used with any other operators as
simply as :

    cat rdbtable.Header rdbtable.Body | operator ...

