Re[2]: Relational Database WITHOUT server?

[Werner Stuerenburg]

Sorry, no idea. Let me think, where this idea leads us.

I would write my own with php or the like. If it is really
simple, why not? For example, in php:

$file = fread($filename);

would read the file into an array by lines.

$ar = explode(',', $file[$i]);

would break the ith line (record) into fields (separated by
commas). Here you have your basic system for plain CSV-style
files. You can fetch each record and display it accordingly.

You can also add records easily. You cannot delete them as
easily, as this requires to write the file new, which may be ok
if the file is small. Otherwise you would use null values for
erasure and check for that before you display them.

You can sort the array for the first field, but not for the
others. You don't have unique constraints, you will have to care
for the right kind of values yourself etc. You cannot count
easily, in fact you cannot query as you will probably want to.

The reason is simple: you don't have a real database system. And
if your application is not absolutely trivial, you will want to
have one. Do you want to care for data consistency yourself? For
example, if somebody reads a record and updates it, another
deletes it?

If you follow this path, you will end up in writing a real
database system on your own. Why not? But chances are, the cost
is really high, and solutions are there. Not the ones you are
looking for, though, and exactly for that reason. The reaosn is
simple: If you need a database, you will run into these
consequences.

Now lets look at the server/tcp/ip question.

Any PC in your network may serve as server. In fact, even in a
self made system, one has to store the file which is used by all
the others. So that is the server which provides the consistent
database, if you call it that way or not.

Those PCs have to communicate somehow. Why not additionally
through TCP/IP? Any Windows PC offers that feature. No problem to
implement. It is just a few clicks in the network applet.

You want to have it simple. I understand that. I learned php and
MySQL as much as I needed it. That was good, as I gained
extremely powerful tools. I can do simple jobs and complex jobs,
as the case may be. It pays to follow the right path without
detour.

Sie schrieben am Montag, 9. Juli 2001, 04:21:45:

> But I'm looking for a SERVERLESS connection.  Either an api or an odbc driver
> that would attach to just the local filesystem and not a server through
> tcp/ip.  For example, the file(s) would reside right below the web directory
> and the cgi would access it directly.




-- 
Herzlich
Werner Stuerenburg            

_________________________________________________
ISIS Verlag, Teut 3, D-32683 Barntrup-Alverdissen
Tel 0(049) 5224-997 407 · Fax 0(049) 5224-997 409
http://pferdezeitung.de


---
!!NEU!!
Fragen und Antworten zu MySQL und dieser Liste unter
-->>  http://www.4t2.com/mysql