fenway

"For editing" is a php issue... does the mysql query pull the desired rows?

No, not without a stored procedure...

To do this properly, you need to generate a table on-the-fly of the desired date periods -- one per interval -- and then join & count as necessary.

I have no idea what you're talking about -- try to explain this again.

Why can't you just do this in mysql? Then it will be truly random.

Do you get an error when using mysql_select_db()?

Within Flash? You sure this is the correct board?

Too many questions all at once... I'm leaning towards MyISAM, but define "huge amounts of text records".

Sure, why not? Can you give an example?

Why not just the mysql order by clause?

Sure, why not?

LOL.

You need to JOIN these two tables together... you can then use PHP to "indent" whenever the id changes.

Why not just WHERE signup_date < CURDATE() - INTERVAL 30 DAY ?

Maybe... is there only one "list" per field? If so, you can search for the "last" closing tag and the first opening tag rather easily.

Bounce rate? What's that?

You were close... but then you omitted the derived table and just tried to keep joining.

OK, I suppose that's ok. Next step is to join all 3 table stogether.

Now I don't know what you're doing... I don't see you using a derived table at all.

I meant created the rows dynamically for any arbitrary date range whatsoever -- you've hard-coded the dates, I presume.

You're skipping steps again.... in order for your query to be robust, you ALWAYS have to use a batch id... otherwise it doesn't make sense to select a non-group by'ed column, like price. Isn't the max batch id going to be different for each currency? And yes, once you have this table, you should be able to sum(c1*c2) in mysql with the above query as a derived table.

I meant dynamically....

Seriously?

I think I remember you trying to optimize this query earlier?

Well, to be sure it's not a caching issue, use "SELECT SQL_NO_CACHE .... FROM ....." What's the EXPLAIN like?