Find the same amounts

Question

Hi All,
I have this sample data:
create table sampleTable(
sample_id number,
sample_fy number,
sample_amt number
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(1, 2003, 25);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(1, 2004, 10);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(1, 2005, 10);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(1, 2006, 46);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(2, 2003, 5);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(2, 2004, 12);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(2, 2005, 10);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(2, 2006, 6);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(3, 2003, 75);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(3, 2004, 13);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(3, 2005, 13);
insert into sampleTable(sample_id, sample_fy, sample_amt)
values(3, 2006, 26);
How can I find the rows with same amounts (amt) for the same sample_id and years between 2004 and 2005?
The result should be:
sample_id
1
3
I try lead, but it is comparing rows from different sample_id as:
SELECT sample_id, sample_fy, sample_amt,
       LEAD(sample_amt, 1, 0) OVER (ORDER BY sample_id, sample_fy) - sample_amt AS my_diff
       FROM sampleTable
       WHERE sample_fy between 2004 and 2005;
Any ideas?
Maybe is another function that I can use.
Thanks!

Answer 1

Hi,
I would use GROUP BY:
SELECT    sample_id
FROM      sampletable
WHERE     sample_fy  IN (2004, 2005)
GROUP BY  sample_id
HAVING    COUNT (DISTINCT sample_fy)  = 2
AND       COUNT (DISTINCT sample_amt) = 1
This does not assume that there will be esxactly 1 row for each combination of sample_id and sample_fy.  A given sample_id may be missing one year or the other (in which case it will not produce any output), or there may be multiple rows per year (in which case, the sample_id will be chosen only if all they are all equal.  I assume sample_amt can not be NULL).
If you want to use LEAD, add a PARTITION BY clause:
SELECT  sample_id, sample_fy, sample_amt,
        LEAD(sample_amt) OVER (
                                PARTITION BY  sample_id    -- This line added
                                ORDER BY      sample_fy    -- No point including sample_id
                             ) - sample_amt AS my_diff
       FROM sampleTable
       WHERE sample_fy between 2004 and 2005;
"PARTITION BY sample_id" means that a separate computation will be done for each group of rows that has the same sample_id.