Here are a couple modifications that I had made to the
code. I have classify these as 'Could Haves' [ rather
than 'Should Haves' or 'Must Haves'.]
(1) Replaced 'ressources' with 'resources' throughout
code and table definitions.
(2) renamed 'katid' to 'table_catid_idx' for clarity in
index definitions. My standard is 'table_column_idx'
for index naming.
(3) renamed all the tables using a '_load' extention.
This separted my production systems from my load/parse
cycle. I'll create a test phase and then table backup,
replace, rename process. This is a standard data
warehousing trick for which keeps your production
systems up while one parses/updates data.
(4) There is not an 'empty table' phase in the code.
This means that the update will append the last cycle
(unless I missed something)
(5) Drop Adult content. This may be tagged via a
separate parameter field in the database or filtered
out during the parse. I'm investigating adding the
filter during the parse phase. I don't recommend
filtering during cusotomer query cycle since one must
use a "LIKE" operator, which in SQL is very slow.
(6) Added code and message to stop splitting cycle.
Recommend moving switch to the config.php file for
splitting step.
(7) I have done significant changes to the INSERT cycle
of the code. Most are just custom modifications for my
production system. But one that is universal is the
addtions of a TRIM step inside the INSERT SQL. I
perform this for all the text type columns since I have
discovered some editors add spaces at the end of thier
entries. I was wondering about the 'addslashes' code.
It seems necessary only in a shorted list of columns
(subject for further investigation.)
(8) Combine link and description into a single table.
After the parsing cycle I combine Link and Description
into a single table. This speeds up my query cycle
since I eliminated a table read. This is another
standard data warehousing trick in which tables are
De-Nomalized to speed customer queries.
These are just some light changes I found helpful for
my code.
James
Logged In: YES
user_id=926217
Some of your ideas/changes have made it in the new version (3.0)
Thanks for your contribution! I really appriciate it.
Kind regards,
Amir