fenway

You should be able to CONVERT to an accent insensitive collation, like latin1.

That's correct -- there's no difference whatsoever, the php function is simply a wrapper.

I considered that but I'm more so concerned with "unique groups" as apposed to unique columns. I may just be better off doing this with two different queries, php and a loop. Unless anyone else has any other ideas? This is so simple yet so complex at the same time. Wait, why won't this work? Make the UNIQUE index span both columns... then use INSERT IGNORE.

Or you can work around this during the INSERT stage and make it really easy.

There's a sticky in the child board with details about this error.

That sounds impossible... unless t1 doesn't have the same column types.

Seems ok.

It's dangerous to rely on mysql to return rows in any order with an order by clause: SELECT * FROM ( SELECT * FROM comments LIMIT 38,10 ) ORDER BY id DESC;

Why not just: SELECT @statefieldid:=fieldid FROM `#__comprofiler_fields` WHERE `name`='cb_state';

What doesn't work now? TLDR.

Why not use the CSV engine?

ENUM is not the same as INT....

Far too little information... what's "data"?

Say what?

Don't mix the primary key with the index on the url field.

Um... no where clause?

You need to narrow that down... comment out most of the procedure, make sure the definition itself works first.

There's nothing wrong with that... not sure what you mean by doubled -- you're asking for *.

Just remember DISTINCT in **NOT** a function.

Maybe you can explain this "list" in more detail?

Then it's the wrong password.

Yes... one join, one join condition.

EAV tables are your friend.

Nope... what statement?

No, it won't -- that's a cartesian join.