Re: use-revolution digest, Vol 1 #1063 - 17 msgs

[Yennie]

Hi,

This is true, but I don't think it really requires a SQL solution.

Since RevDB supports buildling a database cursor (revdb_query), it doesn't make sense to me that it can't return a subset of rows. I can't even ask a cursor for the _current_ row without iterating through all of the columns. On the other hand, revdb_querylist returns rows- but always ALL of them. I can always manually trim the list afterwards, but this is also awfully inefficient in the case where I am forced to load 5,000 records and then trim them down to 50.

Shouldn't there be a simple revdb_getrows() function to go with revdb_getcolumnxxx()? Certainly revdb can do something more efficient than forcing the scripter to iterate columns with a cursor. Imagine fetching rows 1000 to 1100 out of 2000 in a 20 column results. I would have to iterate over 1,000 records, and fetch each of 20 columns 100 times- that's 3,000 calls to revdb!

Even if revdb couldn't do anything much more efficient internally (which I doubt), it would at least run much faster and save the scripter a lot of headaches.

Has anyone had this same problem, or found a reasonable workaround?

Regards,
Brian

Regards,
Brian

Hi Brian,

Unfortunately, ANSI-SQL doesn't have a native solution
for that: you get all the records that fall within the
conditions of the WHERE-clause.

However, MySQL provides an extension to the standard
syntax - 'LIMIT' - which will do what you want. Eg:
� SELECT cust_id, cust_name FROM customer WHERE
� �� cust_id >= theCurrentId LIMIT 100
would get you the first 100 customers with an ID
higher than or equal to the one in theCurrentID.

Other databases _may_ support the same syntax, or
provide their own similar extension. But as they say:
your mileage may vary...

Hope this helped,