Question

Say I want to do this with SQL (Sybase): Find all fields of the record with the latest timestamp.

One way to write that is like this:

select * from data where timestamp = (select max(timestamp) from data)

This is a bit silly because it causes two queries – first to find the max timestamp, and then to find all the data for that timestamp (assume it’s unique, and yes – i do have an index on timestamp). More so it just seems unnecessary because max() has already found the row that I am interested in so looking for it again is wasteful.

Is there a way to directly access fields of the row that max() returns?

Edit: All answers I see are basically clever hacks – I was looking for a syntactic way of doing something like max(field1).field2 to access field2 of the row with max field1

Answer 1

No, using an aggregate means that you are automatically grouping, so there isn’t a single row to get data from even if the group happens to contain a single row.

You can order by the field and get the first row:

set rowcount 1
select * from data order by timestamp desc

(Note that you shouldn’t use select *, but rather specify the fields that you want from the query. That makes the query less sensetive to changes in the database layout.)