UBB.Developers Forums Community Development Programming PHP and Server-side Scripting Data Normalization?

Hello,

I wonder wether it's good to use full database normalization when dealing with flat files databases?

Also, I wonder why does UBB 6 use perl code in some of it's data files? I mean it takes more time because Perl has to compile these files.

Thanks in advance

howdy JB, good to see you around Tho Charles, LK, jordo, or one of the other developers may be better suited to answer you

Perl interprets the files you require (if they are simply hashes written to file) by parsing them, which is essentially what you are doing if you are reading a flat file - parsing it.

So instead of writing your own parsing routines for your own storage format (although memberlist.cgi comes to mind), it seems to me to be better in this case to rely on the existing methods as they are going to be running (according to asymptotic analysis) just as fast.

In terms of doing db normalization on the flat file db, I would guess that Charles probably does do that to some degree when deciding on new storage structures. If you look in the members directory of your 6+ series board, you are going to notice a lot of redundant data (last post histories for members, etc) to make accessing certainly commonly accessed data much faster than reading it from the current structures.

Heh, if you think the indexes in 6.3 are redundant, wait until you see the index files we use in 6.4...

We store serialized Perl data structures rather than extruded data for two reasons:

1) Parsing time. OpenFileAsArray & split()ing results in massive slowdowns. Parsing a Perl data structure is a comparatively lightweight function

2) Simplicity. It's far easier to RequireVars (thus do()) something.

Allen, I've been here for a while
Jordo and Charles, Thank you very much for your explanations