Select last value for each day from table

Question

Hi!
I have a table that stores several measures for each day. I need two queries against this table and I am not quite sure how to write them.
The table stores these lines (sample data)
*DateCol1                 Value       Database*
27.09.2009 12:00:00       100           DB1
27.09.2009 20:00:00       150           DB1
27.09.2009 12:00:00       1000          DB2
27.09.2009 20:00:00       1100          DB2
28.09.2009 12:00:00       200           DB1
28.09.2009 20:00:00       220           DB1
28.09.2009 12:00:00       1500          DB2
28.09.2009 20:00:00       2000          DB2Explanation of data in the sample table:
We measure the size of the data files belonging to each database one or more times each day. The value column shows the size of the database files for each database at a given time (European format for date in DateCol1).
What I need:
Query 1:
The query should return the latest measurement for each day and database. Like this:
*DateCol1       Value      Database*
27.09.2009        150          DB1
27.09.2009       1100          DB2
28.09.2009        220          DB1
28.09.2009       2000          DB2Query 2:
The query should return the average measurement for each day and database. Like this:
*DateCol1       Value      Database*
27.09.2009       125          DB1
27.09.2009      1050          DB2
28.09.2009       210          DB1
28.09.2009      1750          DB2Could someone please help me to write these two queries?
Please let me know if you need further information.
Edited by: user7066552 on Sep 29, 2009 10:17 AM
Edited by: user7066552 on Sep 29, 2009 10:17 AM

Answer 1

For first query you can use analytic function and solve it.
with t
as
select to_date('27.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss') dt,       100 val,           'DB1' db from dual union all
select to_date('27.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       150,           'DB1' from dual union all
select to_date('27.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss'),       1000,          'DB2' from dual union all
select to_date('27.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       1100,          'DB2' from dual union all
select to_date('28.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss'),       200,           'DB1' from dual union all
select to_date('28.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       220,           'DB1' from dual union all
select to_date('28.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss'),       1500,          'DB2' from dual union all
select to_date('28.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       2000,          'DB2' from dual
select dt, val, db
  from (
select row_number() over(partition by trunc(dt), db order by dt) rno,
       count(*) over(partition by trunc(dt), db) cnt,
       t.*
  from t)
where rno = cntFor second you can just group by
with t
as
select to_date('27.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss') dt,       100 val,           'DB1' db from dual union all
select to_date('27.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       150,           'DB1' from dual union all
select to_date('27.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss'),       1000,          'DB2' from dual union all
select to_date('27.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       1100,          'DB2' from dual union all
select to_date('28.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss'),       200,           'DB1' from dual union all
select to_date('28.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       220,           'DB1' from dual union all
select to_date('28.09.2009 12:00:00', 'dd.mm.yyyy hh24:mi:ss'),       1500,          'DB2' from dual union all
select to_date('28.09.2009 20:00:00', 'dd.mm.yyyy hh24:mi:ss'),       2000,          'DB2' from dual
select trunc(dt) dt, avg(val) val, db
  from t
group by trunc(dt), db
order by trunc(dt)