Advice on decode or case

Hi
I'm pretty new to SQL and need some help with SQL syntax.
I am trying to write a sql query which takes as input a variable and dependent upon the value of this variable queries for specific values from the database. E.g.,
if varA =x then query d/b for values y1, y2, y3
elsif varA =c then query d/b for values d1, d2
else blah
end
So, I've attempted to write the sql within a decode but I am getting and error (ANELVA invalid identifier).
(input variable below is vartoolset so I'm saying that if vartoolset=Anelva then I query for ANL_51 & ANL_52 from job_history table)
select facility_code,
decode (vartoolset,"Anelva", (select distinct(facility_code) from job_history where facility_code='ANL_51' OR facility_code='ANL_52'),
"GMR", (select facility_code from job_history where facility_code='CVC_81' OR facility_code='CVC_82'),
'blah')
from job_history
Thanks in advance for any help/advice
Emma

The possible values which may be assigned to
vartoolset are:
Anelva
CVC
GMR
602G
If vartoolset has a value of Anelva then I want to
create a list of the following tools onlys:
ANL_51, ANL_52,ANL_53, ANL_54
If vartoolset has a value of CVC then I want to
create a list of the following tools only:
CVC_51,CVC_52,CVC_54 etc
The tools above are being pulled from column called
facility_code in table job_history.
Does this clarify?
ThanksNot clear..
This?
sql>select * from job_history;
VARTOOLSET FACILITY_CODE
Anelva ANL_51
Anelva ANL_52
Anelva ANL_53
Anelva ANL_54
Anelva CVC_51
CVC CVC_51
CVC CVC_52
CVC CVC_53
CVC CVC_54
CVC ANL_51
GMR ANL_51
GMR CVC_51
SQL>
select facility_code,vartoolset
from job_history
where (vartoolset,facility_code) in (('Anelva','ANL_51'),('Anelva','ANL_52'),('CVC','CVC_51'),('CVC','CVC_52'));
FACILITY_C VARTOOLSET
ANL_51 Anelva
ANL_52 Anelva
CVC_51 CVC
CVC_52 CVC

Similar Messages

Diff between decode and case

Hi
I need small clarification about difference between decode and case
Thanks who visit my thread

And for those people who can't be ar$ed to follow links...
Decode allows for conditional output where the conditions are straightforward matches between values.
Case allows for regular conditions to be used such as greater than, less than etc.
Some of the things done by case can be achieved through decode as well but in a more convoluted manner than simply using case.
;)

Decode Vs Case Statement

What will be good for proformance wise:
Decode or Case in a sql statement.?????

See the following link for Tom Kyte's opinion (point #4 in his first answer):
http://asktom.oracle.com/pls/ask/f?p=4950:8:16717708356827415201::NO::F4950_P8_DISPLAYID,F4950_P8_CRITERIA:1243867216406

Which is the best decode or case

Hi,
When you check performance wise which is the best one decode or case?
Thanks,

> You mean CPU processor speed or oracle buffer(SGA).
Neither. CPU architecture. RISC vs CISC vs ..
On a PA-RISC1 CPU a DECODE is just a tad faster than a CASE. On an AMD64 CPU, the reverse is true.
> When I increase memory, The case and decode performance will increase?
No. A CASE and a DECODE does not need memory to work faster. It is a set of machine code instructions that needs to compare values to determine a result. It depends on just how fast the CPU can execute this set of machine code instructions.
A faster CPU will make a very significant difference. An AMD64 Opteron CPU is a couple of times faster than a PA-RISC1 CPU.
I had this exact same conversation back in 2006 on this forum - and posted [url
http://forums.oracle.com/forums/thread.jspa?messageID=1346165&#1346165]this benchmark to show that the decision of using CASE or DECODE is not a decision that should be based on raw performance.

Without Decode or Case

I have emp_allocation table. It has data as below
EMPID     YEAR      MONTH
X     2006     JAN
X     2006     MAR
Y     2006     JAN
Y     2006     FEB
Y     2006     MAR
I want SQL Query(Without Decode or Case) which will give output as below
EMPID     YEAR      JAN     FEB     MAR      APR     JUN     JUL
X     2006     Y     N     Y
Y     2006     Y     Y     Y

Why you'd want to do it this way I do not know, but if you insist...
SQL> ed
Wrote file afiedt.buf
1 WITH t AS (select 'X' AS EMPID, 2006 AS YEAR, 'JAN' AS MONTH FROM DUAL UNION ALL
2             select 'X', 2006, 'MAR' FROM DUAL UNION ALL
3             select 'Y', 2006, 'JAN' FROM DUAL UNION ALL
4             select 'Y', 2006, 'FEB' FROM DUAL UNION ALL
5             select 'Y', 2006, 'MAR' FROM DUAL)
6 -- END OF TEST DATA
7 SELECT EMPID, YEAR, NVL(MAX(JAN),'N') AS JAN, NVL(MAX(FEB),'N') AS FEB, NVL(MAX(MAR),'N') AS MAR, NVL(MAX(APR),'N') AS APR
8 FROM
9 (
10    SELECT EMPID, YEAR, 'Y' AS JAN, NULL AS FEB, NULL AS MAR, NULL AS APR FROM t WHERE MONTH = 'JAN' UNION ALL
11    SELECT EMPID, YEAR, NULL AS JAN, 'Y' AS FEB, NULL AS MAR, NULL AS APR FROM t WHERE MONTH = 'FEB' UNION ALL
12    SELECT EMPID, YEAR, NULL AS JAN, NULL AS FEB, 'Y' AS MAR, NULL AS APR FROM t WHERE MONTH = 'MAR' UNION ALL
13    SELECT EMPID, YEAR, NULL AS JAN, NULL AS FEB, NULL AS MAR, 'Y' AS APR FROM t WHERE MONTH = 'APR'
14 )
15 GROUP BY EMPID, YEAR
16* ORDER BY 1,2
SQL> /
E       YEAR J F M A
X       2006 Y N Y N
Y       2006 Y Y Y N
SQL>

Decode and Case

Hi,
What is the difference between decode and case?
What are the cases in which we can use Decode and case
Thanx

you can not put Search CASE statements by using DECODE
Eg:
SELECT AVG(CASE WHEN e.salary > 2000 THEN e.salary ELSE 2000 END) "Average Salary" FROM employees e;Can't we?
select avg(decode(sign(e.salary - 2000), 1, e.salary, 2000)) "Average Salary" from employees e;

ANSI Standard decode to CASE

Hi all,
select
id
,one
,two
From
     (select id
     , one
     , two
     , Max(decode(TYPE, "ER", 3*4, 4*7)
     from
          (select id
          , sum(one) as one
          , sum(two) as two
          from t1
          group by id)
     group by id, one, two
     ) temp inner join t2....
The above query is just a sample structure to tell the problem that i am facing. It works fine, but i need to convert it to ANSI Standard.
Hence I need to convert the DECODE to CASE statement. but if i convert to CASE, i need to use any of the GROUP BY FUNCTIONS, but if i use the MAX of any group by functions, the result would be wrong...how to overcome it
Thanks

Hi,
Yes I did a silly mistake, the code I gave was for only the decode but when I converted it to case statement, there was a silly mistake, which i corrected.
Now i corrected it to
Select
     COL1
     , Col2
     , MAX(CASE WHEN Upper(TYPE) = 'YR' THEN Round(3.43*3, 2) END) AS CALC1
     , Max(CASE WHEN Upper(TYPE) = 'ZP' THEN Round(3.12*12, 2) END) AS CALC2
From
     Group by....
but now when i add a sum of the calc1 & calc2, I am not getting any results the column is blank. Is there anything that i missed
Select
     COL1
     , Col2
     , MAX(CASE WHEN Upper(TYPE) = 'YR' THEN Round(3.43*3, 2) END) AS CALC1
     , Max(CASE WHEN Upper(TYPE) = 'ZP' THEN Round(3.12*12, 2) END) AS CALC2
     , (MAX(CASE WHEN Upper(TYPE) = 'YR' THEN Round(3.43*3, 2) END
     + Max(CASE WHEN Upper(TYPE) = 'ZP' THEN Round(3.12*12, 2) END)) as SUM3
From
     Group by....
Thanks

Converting Decode to Case

I'm a beginner with Oracle SQL and I have a select statement with the following code:
max(decode(EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD, '1000', EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE, '2000', EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE, '4000', EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE))
I've been asked to create several more columns in a similar manner. My concern is that since this relies on 'decode' which is less efficient than 'case', that adding more columns using this approach will bog down an already not efficient query. Bottom line my problem is that I don't really understand this 'decode' since all the explanations of 'decode' I've found stop at four parameters.
Could someone please show me how the expression above looks in If-then-else terms as well as comment on how to convert this 'decode' to 'case'? Thanks in advance!

1) The if-then-else stmt
-- Longest way
if EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD = '1000' then
return value of EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE;
elsif EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD = '2000' then
return value of EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE;
elsif EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD = '4000' then
return value of EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE;
else return null;
end if;
-- shorter way
if EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD = '1000' or
EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD = '2000' or
EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD = '4000' then
return value of EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE;
else return null;
end if;
-- shortest if-then-else code
if EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD in ('1000', '2000', '4000') then
return value of EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE;
else return null;
end if;
2) The MAX() function
The MAX function will return the greatest value of EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE in step 1 above.
2) Using CASE
select
max(case when (EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_CD in('1000','2000','4000'))
then EDW.V_RECRUIT_TEST_RESULT_HIST.TEST_RESULT_TEST_SCORE end)
into ...your_item_or_local_variable
from EDW.V_RECRUIT_TEST_RESULT_HIST;
Try the code above, good luck
v/r
Vien Tran

What is better and fast to use between decode and case

Hi friends,
i wanted to know what is better to use decode or case in sql,which gives faster result.
thks
sonal....

Here's a very simple timing comparison. This table (actually it's a partition) has a little over 1 million rows.
As you can see, the timing difference is trivial. (I ran an earlier query to fetch the blocks from disk).
SQL> select sum(decode(balloon_flag
2                   ,'Y',1
3                   ,0
4                   )
5            ) bal_count
6 from   prod.loan_master
7 where report_date = to_date('31-DEC-2005');
           BAL_COUNT
                9036
Elapsed: 00:00:07.65
SQL> select sum(case balloon_flag
2                  when 'Y' then 1
3                  else 0
4             end
5            ) bal_count
6 from   prod.loan_master
7 where report_date = to_date('31-DEC-2005');
           BAL_COUNT
                9036
Elapsed: 00:00:07.68
SQL> select sum(case
2                  when balloon_flag = 'Y' then 1
3                  else 0
4             end
5            ) bal_count
6 from   prod.loan_master
7 where report_date = to_date('31-DEC-2005');
           BAL_COUNT
                9036
Elapsed: 00:00:07.46

How does decode and case works?

Hi,
I want to know how decode and case works? How they process the arguements. Please explain me with some examples.
Thanks,
Sarav
Edited by: 943941 on Jul 3, 2012 1:42 AM

welcome
check this link
https://forums.oracle.com/forums/ann.jspa?annID=1535
you will find everything you need
DECODE compares expr to each search value one by one. If expr is equal to a search, then Oracle Database returns the corresponding result. If no match is found, then Oracle returns default. If default is omitted, then Oracle returns null.
This example decodes the value warehouse_id. If warehouse_id is 1, then the function returns 'Southlake'; if warehouse_id is 2, then it returns 'San Francisco'; and so forth. If warehouse_id is not 1, 2, 3, or 4, then the function returns 'Non domestic'.
SELECT product_id,
       DECODE (warehouse_id, 1, 'Southlake',
                             2, 'San Francisco',
                             3, 'New Jersey',
                             4, 'Seattle',
                                'Non domestic')
       "Location of inventory" FROM inventories
       WHERE product_id < 1775;Edited by: user 007 on Jul 3, 2012 2:40 PM

DECODE OR CASE - Which is better and why

Oracle version: 11.2
Problem: We have a huge table with 10M records, which needs to be processed daily.
While loading data in table we have to handle condition if flag is =1 then '111' else '000'
To implement this which one is efficient solution? .. using CASE or DECODE and why?
Example:
SQL> select flag,case when flag='Y' then '111' else '000' end from (select 'Y' as flag from dual union all select 'N' from dual);
F CAS
Y 111
N 000
SQL> select flag,decode(flag,'Y','111','000') from (select 'Y' as flag from dual union all select 'N' from dual);
F DEC
Y 111
N 000

Hi,
For this job, they're equally efficient.
For any job, where DECODE doesn't require a lot more code than CASE, they will be equally efficient.
I recommend (almost) always using CASE. It's easier to understand and debug, and, if written correctly, will never be slower than DECODE.
The only situation where I use DECODE is for very simple tasks (like the one you posted) where this is used inside a very complicated expression, and the slightly less coding needed for DECODE makes the larger statement easier to read.

DECODE OR CASE

I have several codes that mean a specific thing that I need to measure. For instance HS002 and HS004 = MedNec, KS004, KS005, KS006, KS007,KS008= AdminDen. These codes are housed in one field the Denied Reason Field and It might be better to do a Case Statement to just change the code to fit in the one group and then I can count each group. Any help is appreciated. Thanks!!
Im doing a report in which I need to get a count ot How Many Denied, How Many Approved with Denial Reasons and type. Can I get some direction on what would be the best way to do this. Here is my code so far:
SELECT
A1.AUTH_NUMBER,
A1.AUTH_TYPE,
A1..PLACE_OF_SERVICE,
A1.OVERALL_STATUS,
A1.CLOSED_REASON,
A1.DENIED_REASON,
DECODE (A1.DENIED_REASON,
'HS002', 'MED NECESS',
'HS004', 'MED NECESS')AS MEDNEC,
A1.admit_primary_date,
AP.decision_date,
AP.ADVISOR_DECISION,
PM.LAST_NAME,
PM.PROVIDER_ID,
A1.INSERT_DATETIME,
A1.ACTIVE_PHYSICIAN_ADVISOR,
MV.LINE_OF_BUSINESS
FROM Auth_master a1
INNER JOIN Auth_phys_advisor ap
ON a1.auth_number=ap.auth_number
INNER JOIN Prov_master pm
ON ap.seq_prov_id=pm.seq_prov_id
LEFT JOIN Note_master nm
ON nm.seq_memb_id=a1.seq_memb_id
INNER JOIN Member_mv mv
ON mv.seq_memb_id=a1.seq_memb_id
Where mv.Line_of_Business <>'SFS'
AND A1.PLACE_OF_SERVICE IN ('11','21','22','24')
AND a1.insert_datetime Between To_Date ('04/01/2012', 'MM/DD/YYYY') and To_Date ('04/30/2012','MM/DD/YYYY')

select mydata.mednec, count(*)
from (
SELECT A1.AUTH_NUMBER,
       A1.AUTH_TYPE,
       A1.PLACE_OF_SERVICE,
       A1.OVERALL_STATUS,
       A1.CLOSED_REASON,
       A1.DENIED_REASON,
       Decode (A1.DENIED_REASON, 'HS002', 'MED NECESS',
                                 'HS004', 'MED NECESS',
                                 'KS008', 'ADMIN DEN') AS MEDNEC,
       A1.admit_primary_date,
       AP.decision_date,
       AP.ADVISOR_DECISION,
       PM.LAST_NAME,
       PM.PROVIDER_ID,
       A1.INSERT_DATETIME,
       A1.ACTIVE_PHYSICIAN_ADVISOR,
       MV.LINE_OF_BUSINESS
FROM   Auth_master a1
       inner join Auth_phys_advisor ap
         ON a1.auth_number = ap.auth_number
       inner join Prov_master pm
         ON ap.seq_prov_id = pm.seq_prov_id
       left join Note_master nm
         ON nm.seq_memb_id = a1.seq_memb_id
       inner join Member_mv mv
         ON mv.seq_memb_id = a1.seq_memb_id
WHERE mv.Line_of_Business != 'SFS'
       AND A1.PLACE_OF_SERVICE IN ( '11', '21', '22', '24' )
       AND a1.insert_datetime BETWEEN To_date ('04/01/2012', 'MM/DD/YYYY') AND
                                      To_date (
                                      '04/30/2012', 'MM/DD/YYYY')
) mydata
group by mydata.mednecEdited by: AlanWms on May 10, 2012 12:23 PM Added KS008/Admin den
Edited by: AlanWms on May 10, 2012 12:25 PM

Decode and case statement in the update..

Its is more to it, but I want to check with you guys, the experts on this, this look busy to me, it should be a more simplify way of doing it, do you think will work
The government decide to change the ethnic codes, and we have to come with certain rules to in our report, anyway, do you think this will work? again It is more to it I declare my variables, this is just one part of the precedure.
BEGIN
      UTL_FILE.fclose_all;
      v_file_handle := UTL_FILE.fopen (v_out_path, v_out_file, 'a');
      UTL_FILE.put_line (v_file_handle,
                         CHR (10) || TO_CHAR (SYSDATE, 'DD-MON-YYYY HH:MI:SS')
      UTL_FILE.put_line (v_file_handle, 'Entering spbpers_update');
      SELECT upd_spbpers_upd_cur IS
          spriden_pidm,
          szscapp_birth_state,
          szscapp_birth_city,
          DECODE(szscapp_hisp_or_latino_ind,Y,'2',N,'1'),
          DECODE(szscapp_hisp_or_latino_options,XCM,'2',CUB,'2',MEX,'2',PRI,'2',XSM,'2',ESP,'2',XOH,'2'),
          DECODE(szscapp_amer_indn_alaska_opt,XAN,'1','1',XCW,'1',XCH,'1',XCK,'1',XNV,'1',XSX,'1'),
          DECODE(szscapp_amer_indn_alaska_other,XON,'1') (,IND,'1',JPN,'1',KOR,'1',PAK,'1',PHL,'1',VNM,'1',XEA,'1',XIS,'1',XSA,'1'),
          DECODE(szscapp_asian_options,IND,'1',JPN,'1',KOR,'1',PAK,'1',PHL,'1',VNM,'1',XEA,'1',XIS,'1',XSA,'1'),   ,
          DECODE(szscapp_other_east_asia,(IND,'1',JPN,'1',KOR,'1',PAK,'1',PHL,'1',VNM,'1',XEA,'1',XIS,'1',XSA,'1'),
          DECODE(szscapp_other_indian_subcont,XIS,'1'),
          DECODE(szscapp_other_southeast_asia,XSA,'1'),
          DECODE(szscapp_blk_or_afr_amer_opt,XAA,'1',XAF,'1',XCB,'1',XOA,'1'),
          DECODE(szscapp_blk_or_afr_amer_other,XOA,'1'),
          DECODE(szscapp_natve_hawaian_options,GUM,'1',XHI,'1',ASM,'1',XOP,'1'),
          DECODE(szscapp_hawaiian_other,XOP,'1'),
          DECODE(szscapp_white_options,XEU,'1',XME,'1',XOW,'1'),
          DECODE(szscapp_white_other(XOW,'1')
     FROM
         saturn_midd.szscapp
     WHERE
     spriden_id = szscapp_id
     AND spriden_ntyp_code = 'CAPL'
   IF upd_spbpers_upd_cur%ISOPEN
      THEN
         CLOSE upd_spbpers_upd_cur;
      END IF;
      OPEN upd_spbpers_upd_cur;
      LOOP
         FETCH upd_spbpers_upd_cur
          INTO v_pidm,v_birth_state,v_birth_city,v_latino_ind,v_latino_options,
               v_indn_alaska_opt,v_indn_alaska_other,v_asian_options,
               v_other_east_asia,v_other_indian_subcont,v_other_southeast_asia,
               v_blk_or_afr_amer_opt,v_blk_or_afr_amer_other,v_natve_hawaian_options,
               v_hawaiian_other,v_white_options,v_white_other;
         EXIT WHEN upd_spbpers_upd_cur%NOTFOUND;
         IF upd_spbpers_upd_cur%FOUND
           UPDATE saturn.spbpers
                       set SPBPERS_ETHN_CODE = CASE
               WHEN v_latino_ind            IS NOT NULL THEN (spbpers_ethn_code = v_latino_ind,spbpers_activity_date = sysdate)
               WHEN v_latino_options        IS NOT NULL THEN (spbpers_ethn_code = v_latino_options,spbpers_activity_date = sysdate)
               WHEN v_indn_alaska_opt       IS NOT NULL THEN (spbpers_ethn_code = v_indn_alaska_opt,spbpers_activity_date = sysdate)
               WHEN v_indn_alaska_other     IS NOT NULL THEN (spbpers_ethn_code = v_indn_alaska_other,spbpers_activity_date = sysdate)
               WHEN v_asian_options         IS NOT NULL THEN (spbpers_ethn_code = v_asian_options,spbpers_activity_date = sysdate)
               WHEN v_other_east_asia       IS NOT NULL THEN (spbpers_ethn_code = v_other_east_asia,spbpers_activity_date = sysdate)
               WHEN v_other_indian_subcont IS NOT NULL THEN (spbpers_ethn_code = v_other_indian_subcont,spbpers_activity_date = sysdate)
               WHEN v_other_southeast_asia IS NOT NULL THEN (spbpers_ethn_code = v_other_southeast_asia,spbpers_activity_date = sysdate)
               WHEN v_blk_or_afr_amer_opt   IS NOT NULL THEN (spbpers_ethn_code = v_blk_or_afr_amer_opt,spbpers_activity_date = sysdate)
               WHEN v_blk_or_afr_amer_other IS NOT NULL THEN (spbpers_ethn_code = v_blk_or_afr_amer_other,spbpers_activity_date = sysdate)
               WHEN v_natve_hawaian_options IS NOT NULL THEN (spbpers_ethn_code = v_natve_hawaian_options,spbpers_activity_date = sysdate)
               WHEN v_hawaiian_other        IS NOT NULL THEN (spbpers_ethn_code = v_hawaiian_other,spbpers_activity_date = sysdate)
               WHEN v_white_options         IS NOT NULL THEN (spbpers_ethn_code = v_white_options,spbpers_activity_date = sysdate)
               WHEN v_white_other           IS NOT NULL THEN (spbpers_ethn_code = v_white_other,spbpers_activity_date = sysdate)
               WHEN v_birth_state           IS NOT NULL THEN (spbpers_stat_code_birth = v_birth_state,spbpers_activity_date = sysdate)
               WHEN v_birth_city            IS NOT NULL THEN (spbpers_city_birth = v_birth_city,spbpers_activity_date = sysdate)
               WHERE spbpers_pidm = v_pidm;
              END
                 END IF;
      END LOOP;

Did the procedure compile ?
Doesn't look like a right Decode syntax.
DECODE (col1,'VAL1','This','VAL2','That','ElseThis')
means
--Psuedocode
IF col1 = 'VAL1' THEN 'This'
IF col1 = 'VAL2' THEN 'That'
ELSE 'ElseThis'You can use CASE statement Instead of DECODE
CASE
when      szscapp_amer_indn_alaska_other
     in ('XON','IND','JPN','KOR','PAK' ..... ) THEN '1'
when      szscapp_hisp_or_latino_options
     in ('XCM','CUB','MEX','PRI','XSM','ESP','XOH' ...) THEN '2'
END SS

Query to find out sum by using decode or case

Hi, I have data like below and required an output as given below, could you please help me to get the same by using decode / case .....Thanks for your help
INSNAME INSID STATUS
AAA 1000 Scheduled
AAA 1000 Processing
BBB 1001 Inspector
BBB 1001 Scheduled
BBB 1001 Processing
BBB 1001 Inspector
CCC 1002 Scheduled
CCC 1002 Processing
CCC 1002 Inspector
CCC 1002 Scheduled
CCC 1002 Processing
CCC 1002 Inspector
Required Output...
INSNAME INSID sum_of_scheduled sum_of_Processing sum_of_Inspector
AAA 1000 1 1 0
BBB 1001 1 1 2
CCC 1002 2 2 2

And if you want it with CASE statement:
WITH test_table AS
   SELECT 'AAA' insname, 1000 insid, 'Scheduled'   status FROM DUAL UNION ALL
   SELECT 'AAA' insname, 1000 insid, 'Processing' status FROM DUAL UNION ALL
   SELECT 'BBB' insname, 1001 insid, 'Inspector'   status FROM DUAL UNION ALL
   SELECT 'BBB' insname, 1001 insid, 'Scheduled'   status FROM DUAL UNION ALL
   SELECT 'BBB' insname, 1001 insid, 'Processing' status FROM DUAL UNION ALL
   SELECT 'BBB' insname, 1001 insid, 'Inspector'   status FROM DUAL UNION ALL
   SELECT 'CCC' insname, 1002 insid, 'Scheduled'   status FROM DUAL UNION ALL
   SELECT 'CCC' insname, 1002 insid, 'Processing' status FROM DUAL UNION ALL
   SELECT 'CCC' insname, 1002 insid, 'Inspector'   status FROM DUAL UNION ALL
   SELECT 'CCC' insname, 1002 insid, 'Scheduled'   status FROM DUAL UNION ALL
   SELECT 'CCC' insname, 1002 insid, 'Processing' status FROM DUAL UNION ALL
   SELECT 'CCC' insname, 1002 insid, 'Inspector'   status FROM DUAL
SELECT insname
      ,insid
      ,SUM(CASE WHEN status = 'Scheduled'
                THEN 1
                ELSE 0
            END
          ) sum_of_scheduled
      ,SUM(CASE WHEN status = 'Processing'
                THEN 1
                ELSE 0
            END
          ) sum_of_processing
      ,SUM(CASE WHEN status = 'Inspector'
                THEN 1
                ELSE 0
            END
          ) sum_of_inspector
FROM test_table
GROUP BY insname
        ,insid;
Regards
Arun

Decode or Case when statement

Hi,
I had a condition like
test -table
code - column
t --> True
f --> False
I had two choices of writing query
select decode(code,'t','True','False') from test;
or
select case code
when 't' then 'True'
else 'False'
end
from test;
I want to know which one is best performance wise

Wrong question really.. unless you writing a very tight loop doing 100's of 1000's of iterations using that type of conditional structure. And even then performance difference will not be that significant.
The right question to ask is which one is more flexible and allows for the programmer that comes after you to read and understand and maintain your code. The CASE statement is in this regard, a lot better than a DECODE statement.
Looking just at performance (which I suggest you do not do in isolation!!), I get mixed results using 10G Enterprise on different platforms (HP-UX vs SUN AMD) and operating systems (HP-UX vs Linux).
On the former the DECODE is slightly faster. On the latter the CASE is slightly faster.
Using the construct you've specified and doing a tight loop of a 100,000 iterations, the elapsed execution times are:
HP-UX DECODE = 00:00:11.83
HP-UX CASE = 00:00:12.32
Linux/AMD DECODE =00:00:02.02
Linux/AMD CASE = 00:00:01.84
Obviosuly the CPU architecture plays a major role here. The AMD is considered as the best 64 CPU on the market. The HP-UX PARISC CPU (also 64bit), does not really compare raw performance wise. In addition this is a RISC CPU whereas the AMD CPU is I believe more CISC than RISC.
Also interesting that the faster Sun AMD/Linux server I used for this benchmark is about 10% the price of the HP-UX server.. and about 5x faster ito raw speed as this benchmark showed. :-)
An interesting exercise, but one with little real world value. Yes performance is important. But within PL/SQL, not to the level about debating whether a CASE or DECODE is faster. As I've mentioned, I believe the right question being one about readability and maintenance and not performance in this case.

Advice on decode or case

Similar Messages

Maybe you are looking for