Optimized way of finding the correct tuple.

Question

i have a table which has parameters as
date time value
0 0 q
10/10/07 15:00 k
10/10/06 15:00 U
20/12/08 15:00 R
01/11/09 15:00 I
10/10/09 15:00 E
Data and time are the ranges(start ranges) and there values after that time e.g from (0,0) to ( 10/10/06,15:00) the value will be q and then till (10/10/07 15:00) will be k
I need to find a optimized query in which after supplying data and time , i get the value.
Currently :-
SELECT MAX(date) INTO :ora9_fromdate FROM pap where data<current date;
select FROMDATE, FROMTIME,value from pap where date= ora9_fromdate and time = (select max(time) from papa where date= ora9_fromdate)

Answer 1

Hi,
I'm not sure I understand the problem. Post CREATE TABLE and INSERT statments for the table you posted, CREATE TABLE and INSERT statements for a set of test dates, and the results you want to get for each of those test dates, given the first table.
Why not store the date and time tiogether in on DATE column?
Queries on this table will be more efficient if you can store the next date on each row, so each row of the table has a beginning date and an ending date (actuall, 1 second past the eneding date). That's not hard to compute, using the anlytic LEAD function, but you'd have to do it whenever the table changed (perhaps in a trigger).
How is this table maintained? For example, do you only add new rows with a date after that latest existing date, and rarley update existing rows?
If you can't store the ending date in the table, then what you have now is the best approach. You can make it a little faster by uisng only one query, with the LAST function, like this:
SELECT  MAX (from_dt_tm)                         AS from_dt_tm
,      MIN (value) KEEP (DENSE_RANK LAST ORDER BY from_dt_tm)     AS value
FROM      pap
WHERE      from_dt_tm     <= :target_dt_tm
;