David Blomstrom wrote:
>--- douglass_davis@stripped wrote:
>In my opinion:
>The easiest thing for you to do right now would be a
>ID (primary key) NAME PARENTID
>You don't need PARENT, because you have the PARENTID,
>you can always retrieve parent. Although... i still
>think it would have been simpler to do separate tables
>for each taxonomic level, and this single table design
>may cause problems down the road if you start adding
>more fields... but anyway.
>Keep it simple. One ID is all you need to identify
>any row. As far as the URL, you could use the scheme
>i gave earlier for the URLs (if you can absolutely
>guarantee 100% every NAME-PARENT combination will be
>unique), but it would be more complicated PHP code and
>more complicated queries, and links would break when
>spellings changed. Just use the ID in the URL, it's
>If you add or delete rows, or change spelling, the
>keys stay the same, primary keys don't change. Just
>make sure you don't delete anything's parent and
>everything will be just fine.
>* * * * *
>OK, thanks. I'll give that a try. I'll probably try to
>combine it with the recursive array described at
>(after I learn how to do it).
>Let me throw one more curve ball at you, though. I now
>understand that I can't combine my primary key with
>the field Name, like 8leo. But do you know if I could
>combine numerals from this "tree traversal" script
>Here's why numerals appeal to me. If I combine fields
>like Name/Parent ID - e.g. leo/Pan, then it will be
>more difficult when I want to separate them. If I
>combine Name with a numeral - e.g. leo/8 - then all I
>need is a script that weeds out all numerals. I think.
>Does this make sense?
>Do You Yahoo!?
>Tired of spam? Yahoo! Mail has the best spam protection around
All you need is one ID to identify anything in there. No need to
combine anything . Keep it simple.