fenway

Not true.

I don't understand what you mean.

Sorry... don't have much experience with that.

Sounds like you need to build the where clause on the fly.

At some point you'll hit the FS limit on files.... but you'll hit other practical limits way before them.

I don't understand.

Are those single-quotes?

Retrieved in batches? Somewhere the data was entered one at a time...

Yes, I understand -- so could you explain this in plain English?

That option wasn't clear from the query.

So you're inserting records into this "additional table" as you go along? I'm confused... why not just timestamp them?

Why not (language_id, sort )? And why the left joins?

Like I said, let's see what those two expressions are equal to.

My guess? PHP is having trouble parsing that... use curlies.

Forget about the field for zip codes -- make another table, and make one record per zipcode/company_id pair.

How can Barand's not work? Put those two expressions in the select column list and check them

How can you do this currently?

Then could you explain what you're trying to do, and post the table structures?

You want 10 from each?

Can we see the insert query? Is it the first TIMESTAMP column?

Why do you need a new table? Why not just JOIN the table to itself?

Sorry, I misread your original post, I thought you wanted them all weighted equally. The problem is that SQL doesn't really have a mechanism for "extrapolating" rows in a way that would make this easy. Depending on how "random" you want it to be, you can multiply the number of posts by RAND(), and pick the highest value...

"Split the information"? I really hope you're not storing the date as three separate fields....

So you want an UPDATE statement with a WHERE clause that matches the ID of the record.

So you're saying you've run this query from PHPMyAdmin and it returns records, but not from your script? Also, MySQL can be very lenient in storing date/times.