Pure Perl SQL server

[Peter Cooper]

Does anyone know of a pure Perl SQL server? I have noticed the rise of 'pure
Perl' modules for things like XML parsing, and the pure Perl DBD::MySQLPP,
but have never seen anything as ambition as a pure Perl SQL server.

I am not entirely sure as to how useful such a thing would be, although it'd
make a cute part time project, as there's bound to be tons of things I'd
learn along the way. With Perl6/Parrot approaching rapidly, I believe such
pure Perl systems could come into their own, as Perl's speed rockets up the
y-axis.

This is just blustering at this point, but I'd be interested to hear what
other people think. My searches for such a project have been fruitless,
except for a post on a LUG mailing list where someone was asking for a pure
Perl SQL server, but that was four years ago, and he got no replies! :-)

I have looked at SQL::Statement which does a little parsing for DBD::CSV,
although I have yet to take a good look into DBD::CSV itself. I believe
there's potential to developing a 'reasonably' powerful SQL engine in pure
Perl, and if it's developed with Perl 6 in mind, it may become a rather
useful tool in the future.

Thoughts? Hisses? Sarky comments? Insights? :-)

Regards,
Peter Cooper

[Bob Walton]

Peter Cooper wrote:

> Does anyone know of a pure Perl SQL server? I have noticed the rise of 'pure
> Perl' modules for things like XML parsing, and the pure Perl DBD::MySQLPP,
> but have never seen anything as ambition as a pure Perl SQL server.
>
> I am not entirely sure as to how useful such a thing would be, although it'd
> make a cute part time project, as there's bound to be tons of things I'd
> learn along the way. With Perl6/Parrot approaching rapidly, I believe such
> pure Perl systems could come into their own, as Perl's speed rockets up the
> y-axis.

....

> I have looked at SQL::Statement which does a little parsing for DBD::CSV,
> although I have yet to take a good look into DBD::CSV itself. I believe
> there's potential to developing a 'reasonably' powerful SQL engine in pure
> Perl, and if it's developed with Perl 6 in mind, it may become a rather
> useful tool in the future.

....

> Peter Cooper

Hmmmmm...well, if you use DBI::PurePerl in conjunction with DBD::CSV, it
looks like you probably get a pure Perl SQL database implementation.
There is also DBD::RAM, which also appears to be pure Perl. I'm not
sure just what you mean by "SQL server".

--
Bob Walton

[Peter Cooper]

Thanks for replying, Bob.

> Hmmmmm...well, if you use DBI::PurePerl in conjunction with DBD::CSV, it
> looks like you probably get a pure Perl SQL database implementation.
> There is also DBD::RAM, which also appears to be pure Perl. I'm not
> sure just what you mean by "SQL server".

Good points. I wanted someone to question my idea in this way, so I can at
least attempt to prove/disprove its validity in my head. I didn't know about
DBD::RAM before, but I've checked that out, and it seems interesting.. I'll
have to run a few 'experiments' with it :-)

By "SQL server", I mean replicating something like MySQL entirely in Perl.
So, yeah, you'd create a DBD, and use DBI, as you would with anything else,
but the actual database /backend/ (like your MySQL, Oracle, etc) would be
Perl code.. and it wouldn't be CSV, or in RAM.. but fast binary files.

The featureset would be comparable to an early MySQL, better than
SQL::Statement can parse (and therefore better than DBD::CSV or DBD::RAM can
handle), and using binary files. It'd be no speed champion, but when Perl 6
is out and about, it could be 'reasonably' fast, thanks to Parrot, and it'd
be entirely interpreted. For people who can't even create sockets (for
security purposes), you could also make it run in a 'command line' form.

Just pipedreams, I guess, but I'll be sure to look into it further. Any
further queries or comments would be most appreciated though.

Pete

[Ted Zlatanov]

On Sat, 23 Aug 2003, [email]newsfeed@boog.co.uk[/email] wrote:

> By "SQL server", I mean replicating something like MySQL entirely in
> Perl. So, yeah, you'd create a DBD, and use DBI, as you would with
> anything else, but the actual database /backend/ (like your MySQL,
> Oracle, etc) would be Perl code.. and it wouldn't be CSV, or in
> RAM.. but fast binary files.

That's an interesting idea, but a RDBMS is a very difficult piece of
software to write correctly. I haven't heard about something written
in Perl. I would love to see one, though.

Ted