PERCENTILE_CONT with non constants

Question

Hi All,
I'm currently learning PL/SQL and I'm having an issue with PERCENTILE_CONT where I want to pass a variable value instead of a constant. Can anyone shed any light on how I could coerce the variables into a constant yet still keep it dynamic or alternatively suggest any alternative functions that may perform the percentile calculations?
My SQL is (very ameturish, improvements appreciated) below:
SELECT
     TB1.FRONTEND_COUNTRY,ROUND(AVG(TB1.ER_RT),2) ER_RATE,ROUND(AVG(SFR_RT),2) SFR_RATE,ROUND(PERCENTILE_CONT(TB2.PERC1/100) WITHIN GROUP (ORDER BY TB1.SSV_SIV ASC)*30.5,0) SSV_SIV1, ROUND(PERCENTILE_CONT(TB2.PERC2/100) WITHIN GROUP (ORDER BY TB1.SSV_SIV ASC)*30.5,0) SSV_SIV2
FROM     
     (SELECT PD.FRONTEND_COUNTRY, PD.PROJECT_CODE, PD.PROTOCOL, AVG(PD.PROJECT_MEAN_ERMTHS) ER_RT, AVG(PD.PROJECT_MEAN_SFR_PERCENT) SFR_RT, AVG(PD.COUNTRY_MEAN_SSU2) SSV_SIV
     FROM MYDB.TCTM_TEMP_USERACC_PROJDATA1 PD
     GROUP BY PD.FRONTEND_COUNTRY, PD.PROJECT_CODE, PD.PROTOCOL
     ORDER BY 1) TB1
     LEFT JOIN MYDB.TCTM_STRAT_NEWVAROFMIKE TB2 ON TB1.FRONTEND_COUNTRY=TB2.COUNTRY_NAME
GROUP BY
     TB1.FRONTEND_COUNTRY
ORDER BY 1Edited by: 955750 on 29-Aug-2012 05:53
Edited by: 955750 on 29-Aug-2012 07:12

Answer 1

Hi,
Welcome to the forum!
Whenever you have a problem, please post a little sample data (CREATE TABLE and INSERT statements, relevant columns only) from all tables involved.
Also post the results you want from that data, and an explanation of how you get those results from that data, with specific examples.
It's good that you posted your code; always do that.
Always say which version of Oracle you're using.
See the forum FAQ {message:id=9360002}
955750 wrote:
Hi All,
I'm currently learning PL/SQL Do you mean SQL?
and I'm having an issue with PERCENTILE_CONT where I want to pass a variable value instead of a constant. Can anyone shed any light on how I could coerce the variables into a constant yet still keep it dynamic or alternatively suggest any alternative functions that may perform the percentile calculations?You can pass variables to PERCENTILE_CONT; the question is which values of the variables?
>
My SQL is (very ameturish, improvements appreciated) below:Your SQL isn't bad.
It's hard to read if it's not formatted. When posting on this site, type these 6 characters
\(lower-case only, inside curly brackets) before and after sections of formatted text to preserve spacing.
SELECT
     TB1.FRONTEND_COUNTRY,ROUND(AVG(TB1.ER_RT),2) ER_RATE,ROUND(AVG(SFR_RT),2) SFR_RATE,ROUND(PERCENTILE_CONT(TB2.PERC1/100) WITHIN GROUP (ORDER BY TB1.SSV_SIV ASC)*30.5,0) SSV_SIV1, ROUND(PERCENTILE_CONT(TB2.PERC2/100) WITHIN GROUP (ORDER BY TB1.SSV_SIV ASC)*30.5,0) SSV_SIV2
FROM     
     (SELECT PD.FRONTEND_COUNTRY, PD.PROJECT_CODE, PD.PROTOCOL, AVG(PD.PROJECT_MEAN_ERMTHS) ER_RT, AVG(PD.PROJECT_MEAN_SFR_PERCENT) SFR_RT, AVG(PD.COUNTRY_MEAN_SSU2) SSV_SIV
     FROM MYDB.TCTM_TEMP_USERACC_PROJDATA1 PD
     GROUP BY PD.FRONTEND_COUNTRY, PD.PROJECT_CODE, PD.PROTOCOL
     ORDER BY 1) TB1There's no reason to use an ORDER BY clause in this sub-query.
     LEFT JOIN MYDB.TCTM_STRAT_NEWVAROFMIKE TB2 ON TB1.FRONTEND_COUNTRY=TB2.COUNTRY_NAME
GROUP BY
     TB1.FRONTEND_COUNTRY
ORDER BY 1
Edited by: 955750 on 29-Aug-2012 05:53This problem concerns GROUP BY more that PERCENTILE_CONT.
In the main query, you're GROUPing BY frontend_country, so there will only be 1 row of ouptput for each distinct value of frontend_country.  Let's say, for a particular frontend_country, there are 3 rows, and thy have perc1 values of 50, 50 and 80.  When you computePERCENTILE_CONT(TB2.PERC1/100) WITHIN GROUP (ORDER BY TB1.SSV_SIV ASC)
for that row, which value of perc1 should be used?  50? 80? Something in between?  And why should it use that value? For example, if you want it to use 50, is that because 50 is the lowest value? The most common value?  Some other reason?
If you know that each distinct value of frontend_coutry will only have 1 distinct value of perc1 (and perc2) in the main query, then you can add perc1 (and perc2) to the main query's GROUP BY clause.
Whatever you want to do, there's probably some way to do it.  However, it's unclear what you want to do.  That's why it's so important to post a little sample data and the results you want from that data.