Sangam 19 - Analytic SQL
0 likes214 views
The document discusses SQL analytics and its functions for data analysis, including SQL commands for selecting employee data, ranking, and aggregating salaries. It provides detailed examples of how to use ranking functions such as RANK(), DENSE_RANK(), CUME_DIST(), and PERCENT_RANK() to analyze employee hiring sequences and salary distributions. Additionally, the document outlines the use of window functions to calculate running totals by employee names within departments.
![SQL> select deptno , rtrim(ename,',') enames
2 from ( select deptno,ename,rn
3 from emp
4 model
5 partition by (deptno)
6 dimension by (
7 row_number() over
8 (partition by deptno order by ename) rn
9 )
10 measures (cast(ename as varchar2(40)) ename)
11 rules
12 ( ename[any]
13 order by rn desc = ename[cv()]||','||ename[cv()+1])
14 )
15 where rn = 1
16 order by deptno;
DEPTNO ENAMES
---------- ----------------------------------------
10 CLARK,KING,MILLER
20 ADAMS,FORD,JONES,SCOTT,SMITH
30 ALLEN,BLAKE,JAMES,MARTIN,TURNER,WARD
- Rob Van Wijk](https://image.slidesharecdn.com/sangamanalytics2-191211041448/85/Sangam-19-Analytic-SQL-150-320.jpg)
Sangam 19 - Analytic SQL
- 2. The power of SQL analytics Connor McDonald Database Advocate Copyright © 2019 Oracle and/or its affiliates.
- 3. 3 3
- 4. 4 4
- 5. 5 Me youtube bit.ly/youtube-connor blog connor-mcdonald.com twitter @connor_mc_d 400+ posts mainly on database & development 250 technical videos, new uploads every week rants and raves on tech and the world :-)
- 6. 6 etc... facebook bit.ly/facebook-connor linkedin bit.ly/linkedin-connor instagram bit.ly/instagram-connor slideshare bit.ly/slideshare-connor
- 7. 7 the beauty of analytics
- 9. 9 <function> ( <arg>,<arg>,… ) OVER ( <partition clause> <sorting clause> <windowing clause> )
- 10. 10 <function> ( <arg>,<arg>,… ) OVER | KEEP ( <partition clause> <sorting clause> <windowing clause> )
- 11. 11 that’s it !
- 15. SQL> select empno, ename, job, hiredate, sal 2 from emp 3 order by sal; EMPNO ENAME JOB HIREDATE SAL ---------- ---------- --------- --------- ---------- 7369 SMITH CLERK 17-DEC-80 800 7900 JAMES CLERK 03-DEC-81 950 7876 ADAMS CLERK 12-JAN-83 1100 7521 WARD SALESMAN 22-FEB-81 1250 7654 MARTIN SALESMAN 28-SEP-81 1250 7934 MILLER CLERK 23-JAN-82 1300 7844 TURNER SALESMAN 08-SEP-81 1500 7499 ALLEN SALESMAN 20-FEB-81 1600 7782 CLARK MANAGER 09-JUN-81 2450 7698 BLAKE MANAGER 01-MAY-81 2850 7566 JONES MANAGER 02-APR-81 2975 7902 FORD ANALYST 03-DEC-81 3000 7788 SCOTT ANALYST 09-DEC-82 3000 7839 KING PRESIDENT 17-NOV-81 5000
- 16. 16 "Show me the hiring sequence"
- 17. EMPNO ENAME JOB HIREDATE SAL HIRE_SEQ ---------- ---------- --------- --------- ---------- ---------- 7369 SMITH CLERK 17-DEC-80 800 1 7900 JAMES CLERK 03-DEC-81 950 10 7876 ADAMS CLERK 12-JAN-83 1100 14 7521 WARD SALESMAN 22-FEB-81 1250 3 7654 MARTIN SALESMAN 28-SEP-81 1250 8 7934 MILLER CLERK 23-JAN-82 1300 12 7844 TURNER SALESMAN 08-SEP-81 1500 7 7499 ALLEN SALESMAN 20-FEB-81 1600 2 7782 CLARK MANAGER 09-JUN-81 2450 6 7698 BLAKE MANAGER 01-MAY-81 2850 5 7566 JONES MANAGER 02-APR-81 2975 4 7902 FORD ANALYST 03-DEC-81 3000 10 7788 SCOTT ANALYST 09-DEC-82 3000 13 7839 KING PRESIDENT 17-NOV-81 5000 9 SMITH was hired "first" ADAMS was hired “last"
- 19. SQL> select e.empno, e.ename, e.job, 2 e.hiredate, e.sal, x.seq 3 from emp e, 4 ( select e2.empno, count(*) seq 5 from emp e1, emp e2 6 where e1.hiredate <= e2.hiredate 7 group by e2.empno 8 ) x 9 where e.empno = x.empno 10 order by sal;
- 20. ------------------------------------------------------ | Id | Operation | Name | Rows | Bytes | ------------------------------------------------------ | 0 | SELECT STATEMENT | | 10 | 1390 | | 1 | SORT GROUP BY | | 10 | 1390 | |* 2 | HASH JOIN | | 10 | 1390 | | 3 | MERGE JOIN | | 10 | 310 | | 4 | SORT JOIN | | 14 | 126 | | 5 | TABLE ACCESS FULL| EMP | 14 | 126 | |* 6 | SORT JOIN | | 14 | 308 | | 7 | TABLE ACCESS FULL| EMP | 14 | 308 | | 8 | TABLE ACCESS FULL | EMP | 14 | 1512 | ------------------------------------------------------
- 21. SQL> select empno, ename, job, hiredate, sal, 2 rank() OVER (order by hiredate) as hire_seq 3 from emp 4 order by sal; EMPNO ENAME JOB HIREDATE SAL HIRE_SEQ ---------- ---------- --------- --------- ---------- ---------- 7369 SMITH CLERK 17-DEC-80 800 1 7900 JAMES CLERK 03-DEC-81 950 10 7876 ADAMS CLERK 12-JAN-83 1100 14 7521 WARD SALESMAN 22-FEB-81 1250 3 7654 MARTIN SALESMAN 28-SEP-81 1250 8 7934 MILLER CLERK 23-JAN-82 1300 12 7844 TURNER SALESMAN 08-SEP-81 1500 7 7499 ALLEN SALESMAN 20-FEB-81 1600 2 7782 CLARK MANAGER 09-JUN-81 2450 6 7698 BLAKE MANAGER 01-MAY-81 2850 5 7566 JONES MANAGER 02-APR-81 2975 4 7902 FORD ANALYST 03-DEC-81 3000 10 7788 SCOTT ANALYST 09-DEC-82 3000 13 7839 KING PRESIDENT 17-NOV-81 5000 9
- 22. rank() OVER ( order by hire_date) as hire_seq function sorting clause
- 24. RANK 1, 2, 3, 3, 5, ... functions for ranking
- 25. RANK 1, 2, 3, 3, 5, ... DENSE_RANK 1, 2, 3, 3, 4, ... functions for ranking
- 26. functions for ranking RANK 1, 2, 3, 3, 5, ... DENSE_RANK 1, 2, 3, 3, 4, ... CUME_DIST SQL> select ename, sal, 2 100*cume_dist() over ( order by sal ) as pct 3 from emp 4 order by sal; ENAME SAL PCT ---------- ---------- ------- SMITH 800 7.14 JAMES 950 14.29 ADAMS 1100 21.43 WARD 1250 35.71 MARTIN 1250 35.71 MILLER 1300 42.86 TURNER 1500 50.00 ALLEN 1600 57.14 CLARK 2450 64.29 BLAKE 2850 71.43 JONES 2975 78.57 FORD 3000 92.86 SCOTT 3000 92.86 KING 5000 100.00
- 27. functions for ranking RANK 1, 2, 3, 3, 5, ... DENSE_RANK 1, 2, 3, 3, 4, ... CUME_DIST PERCENT_RANK SQL> select ename, sal, 2 100*percent_rank() over ( order by sal ) pct 3 from emp 4 order by ename; ENAME SAL PCT ---------- ---------- ------- ADAMS 1100 15.38 ALLEN 1600 53.85 BLAKE 2850 69.23 CLARK 2450 61.54 FORD 3000 84.62 JAMES 950 7.69 JONES 2975 76.92 KING 5000 100.00 MARTIN 1250 23.08 MILLER 1300 38.46 SCOTT 3000 84.62 SMITH 800 .00 TURNER 1500 46.15 WARD 1250 23.08
- 28. functions for ranking RANK 1, 2, 3, 3, 5, ... DENSE_RANK 1, 2, 3, 3, 4, ... CUME_DIST PERCENT_RANK NTILE SQL> select ename, sal, 2 ntile(4) over ( order by sal ) as quartile 3 from emp 4 order by ename; ENAME SAL QUARTILE ---------- ---------- ---------- ADAMS 1100 1 ALLEN 1600 2 BLAKE 2850 3 CLARK 2450 3 FORD 3000 4 JAMES 950 1 JONES 2975 3 KING 5000 4 MARTIN 1250 2 MILLER 1300 2 SCOTT 3000 4 SMITH 800 1 TURNER 1500 2 WARD 1250 1
- 29. RANK 1, 2, 3, 3, 5, ... DENSE_RANK 1, 2, 3, 3, 4, ... CUME_DIST PERCENT_RANK NTILE ROW_NUMBER 1, 2, 3, 4, 5, ... functions for ranking
- 33. SQL> select deptno, empno, ename, job, sal 2 from emp 3 order by deptno, empno; DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7369 SMITH CLERK 800 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 30 7499 ALLEN SALESMAN 1600 30 7521 WARD SALESMAN 1250 30 7654 MARTIN SALESMAN 1250 30 7698 BLAKE MANAGER 2850 30 7844 TURNER SALESMAN 1500 30 7900 JAMES CLERK 950
- 34. 34 "department salaries, running total by employee name"
- 35. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, ename;
- 36. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, ename;
- 37. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, ename;
- 38. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, ename;
- 39. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, ename;
- 40. sum(sal) OVER ( partition by deptno order by ename) as running_total function sorting clause partition clause
- 41. 41 be careful
- 42. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, ename;
- 43. SQL> select deptno, empno, ename, job, sal 2 from emp 3 order by deptno, hiredate; DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7369 SMITH CLERK 800 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 30 7499 ALLEN SALESMAN 1600 30 7521 WARD SALESMAN 1250 30 7654 MARTIN SALESMAN 1250 30 7698 BLAKE MANAGER 2850 30 7844 TURNER SALESMAN 1500 30 7900 JAMES CLERK 950 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, hiredate RUNNING_TOTAL ------------- 2450 7450 8750 10875 7075 10075 1100 4100 1600 9400 6650 4450 8150 5400
- 44. 44 lots of power
- 45. SQL> select deptno, job, ename, sal, 2 sum(sal) over () total_sal, 3 sum(sal) over 4 ( partition by deptno) sal_by_dept, 5 sum(sal) over 6 ( partition by deptno, job) sal_by_dept_job 7 from emp 8 order by deptno, job; DEPTNO JOB ENAME SAL ---------- --------- ---------- ---------- 10 CLERK MILLER 1300 10 MANAGER CLARK 2450 10 PRESIDENT KING 5000 20 ANALYST SCOTT 3000 20 ANALYST FORD 3000 20 CLERK ADAMS 1100 20 CLERK SMITH 800 20 MANAGER JONES 2975 30 CLERK JAMES 950 30 MANAGER BLAKE 2850 30 SALESMAN TURNER 1500 30 SALESMAN MARTIN 1250 30 SALESMAN WARD 1250 30 SALESMAN ALLEN 1600 TOTAL_SAL ---------- 29025 29025 29025 29025 29025 29025 29025 29025 29025 29025 29025 29025 29025 29025 SAL_BY_DEPT ----------- 8750 8750 8750 10875 10875 10875 10875 10875 9400 9400 9400 9400 9400 9400 SAL_BY_DEPT_JOB --------------- 1300 2450 5000 6000 6000 1900 1900 2975 950 2850 5600 5600 5600 5600
- 46. and a whole lot more... COLLECT CORR COVAR_POP COVAR_SAMP GROUP_ID GROUPING GROUPING_ID MEDIAN PERCENTILE_CONT PERCENTILE_DISC REGR_ ... STATS_BINOMIAL_TEST STATS_KS_TEST STATS_MODE STATS_MW_TEST STATS_ONE_WAY_ANOVA STATS_F_TEST STATS_CROSSTAB STATS_T_TEST_... STATS_WSR_TEST STDDEV STDDEV_POP STDDEV_SAMP VAR_POP VAR_SAMP VARIANCE
- 47. 47 there is a cost
- 48. 48 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( ... ), 3 max(hiredate) over ( ... ), 4 stddev(commission) over ( .... ) ... ... ... 25 from emp
- 49. 49 a lot of work
- 50. 50 and a little surprise...
- 51. 51 SQL> select deptno, empno, ename, job, sal 2 from emp 3 order by deptno, empno; ---------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| ---------------------------------------------------------------- | 0 | SELECT STATEMENT | | 12 | 1044 | 4 (25)| | 1 | SORT ORDER BY | | 12 | 1044 | 4 (25)| | 2 | TABLE ACCESS FULL| EMP | 12 | 1044 | 3 (0)| ----------------------------------------------------------------
- 52. 52 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp 6 order by deptno, empno ---------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| ---------------------------------------------------------------- | 0 | SELECT STATEMENT | | 12 | 468 | 3 (0)| | 1 | WINDOW SORT | | 12 | 468 | 3 (0)| | 2 | TABLE ACCESS FULL| EMP | 12 | 468 | 3 (0)| ----------------------------------------------------------------
- 53. 53 some possible sorting anomalies
- 54. MOS workarea_size_policy _smm_isort_cap _smm_max_size _newsort_enabled _smm_auto_min_io_size _smm_auto_max_io_size Current versions should all be fine
- 56. 56 aggregations
- 59. 59 SQL> select deptno, sum(sal) 2 from emp 3 group by deptno; DEPTNO SUM(SAL) ---------- ---------- 10 8750 20 10875 30 9400
- 60. 60 SQL> select ename, deptno, 2 sum(sal) over 3 ( partition by deptno) as deptsal 4 from emp 5 order by deptno; ENAME DEPTNO DEPTSAL ---------- ---------- ---------- CLARK 10 8750 KING 10 8750 MILLER 10 8750 JONES 20 10875 FORD 20 10875 ADAMS 20 10875 SMITH 20 10875 SCOTT 20 10875 WARD 30 9400 TURNER 30 9400 ALLEN 30 9400 JAMES 30 9400 BLAKE 30 9400 MARTIN 30 9400 still 14 rows !!!
- 61. 61 two kinds
- 62. 62 aggregation
- 63. 63 reporting aggregation "same aggregate for each row in a partition"
- 64. 64 SQL> select ename, deptno, 2 sum(sal) over ( partition by deptno) as deptsal 3 from emp 4 order by deptno; ENAME DEPTNO DEPTSAL ---------- ---------- ---------- CLARK 10 8750 KING 10 8750 MILLER 10 8750 JONES 20 10875 FORD 20 10875 ADAMS 20 10875 SMITH 20 10875 SCOTT 20 10875 WARD 30 9400 TURNER 30 9400 ALLEN 30 9400 JAMES 30 9400 BLAKE 30 9400 MARTIN 30 9400 same for each row in partition
- 65. 65 windowing aggregation "changing aggregate for each row in a partition"
- 66. <function> ( <arg>,<arg>,… ) OVER ( <partition clause> <sorting clause> <windowing clause> ) defines how broadly the aggregating function applies
- 67. 67 "show me a salary running total"
- 68. SQL> select 2 empno, ename, sal, 3 sum(sal) 4 over ( order by empno 5 rows between unbounded preceding and current row ) as cumtot 6 from emp 7 order by empno; EMPNO ENAME SAL CUMTOT ---------- ---------- ---------- ---------- 7369 SMITH 800 800 7499 ALLEN 1600 2400 7521 WARD 1250 3650 7566 JONES 2975 6625 7654 MARTIN 1250 7875 7698 BLAKE 2850 10725 7782 CLARK 2450 13175 7788 SCOTT 3000 16175 7839 KING 5000 21175 7844 TURNER 1500 22675 7876 ADAMS 1100 23775 7900 JAMES 950 24725 7902 FORD 3000 27725 7934 MILLER 1300 29025
- 69. SQL> select 2 empno, ename, sal, 3 sum(sal) 4 over ( order by empno 5 rows between unbounded preceding and current row ) as cumtot 6 from emp 7 order by empno; EMPNO ENAME SAL CUMTOT ---------- ---------- ---------- ---------- 7369 SMITH 800 800 7499 ALLEN 1600 2400 7521 WARD 1250 3650 7566 JONES 2975 6625 7654 MARTIN 1250 7875 7698 BLAKE 2850 10725 7782 CLARK 2450 13175 7788 SCOTT 3000 16175 7839 KING 5000 21175 7844 TURNER 1500 22675 7876 ADAMS 1100 23775 7900 JAMES 950 24725 7902 FORD 3000 27725 7934 MILLER 1300 29025
- 70. SQL> select 2 empno, ename, sal, 3 sum(sal) 4 over ( order by empno 5 rows between unbounded preceding and current row ) as cumtot 6 from emp 7 order by empno; EMPNO ENAME SAL CUMTOT ---------- ---------- ---------- ---------- 7369 SMITH 800 800 7499 ALLEN 1600 2400 7521 WARD 1250 3650 7566 JONES 2975 6625 7654 MARTIN 1250 7875 7698 BLAKE 2850 10725 7782 CLARK 2450 13175 7788 SCOTT 3000 16175 7839 KING 5000 21175 7844 TURNER 1500 22675 7876 ADAMS 1100 23775 7900 JAMES 950 24725 7902 FORD 3000 27725 7934 MILLER 1300 29025
- 71. SQL> select 2 empno, ename, sal, 3 sum(sal) 4 over ( order by empno 5 rows between unbounded preceding and current row ) as cumtot 6 from emp 7 order by empno; EMPNO ENAME SAL CUMTOT ---------- ---------- ---------- ---------- 7369 SMITH 800 800 7499 ALLEN 1600 2400 7521 WARD 1250 3650 7566 JONES 2975 6625 7654 MARTIN 1250 7875 7698 BLAKE 2850 10725 7782 CLARK 2450 13175 7788 SCOTT 3000 16175 7839 KING 5000 21175 7844 TURNER 1500 22675 7876 ADAMS 1100 23775 7900 JAMES 950 24725 7902 FORD 3000 27725 7934 MILLER 1300 29025
- 72. 72 "give me the sum across 3 rows, ie, one before and one after"
- 73. SQL> select deptno, ename, hiredate, sal, 2 sum(sal) over ( 3 partition by deptno 4 order by hiredate 5 rows between 1 preceding and 1 following) as x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- ---------- 10 CLARK 09-JUN-81 2450 7450 10 KING 17-NOV-81 5000 8750 10 MILLER 23-JAN-82 1300 6300 20 SMITH 17-DEC-80 800 3775 20 JONES 02-APR-81 2975 6775 20 FORD 03-DEC-81 3000 8975 20 SCOTT 09-DEC-82 3000 7100 20 ADAMS 12-JAN-83 1100 4100 30 ALLEN 20-FEB-81 1600 2850 30 WARD 22-FEB-81 1250 5700 30 BLAKE 01-MAY-81 2850 5600 30 TURNER 08-SEP-81 1500 5600 30 MARTIN 28-SEP-81 1250 3700 30 JAMES 03-DEC-81 950 2200
- 74. SQL> select deptno, ename, hiredate, sal, 2 sum(sal) over ( 3 partition by deptno 4 order by hiredate 5 rows between 1 preceding and 1 following) as x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- ---------- 10 CLARK 09-JUN-81 2450 7450 10 KING 17-NOV-81 5000 8750 10 MILLER 23-JAN-82 1300 6300 20 SMITH 17-DEC-80 800 3775 20 JONES 02-APR-81 2975 6775 20 FORD 03-DEC-81 3000 8975 20 SCOTT 09-DEC-82 3000 7100 20 ADAMS 12-JAN-83 1100 4100 30 ALLEN 20-FEB-81 1600 2850 30 WARD 22-FEB-81 1250 5700 30 BLAKE 01-MAY-81 2850 5600 30 TURNER 08-SEP-81 1500 5600 30 MARTIN 28-SEP-81 1250 3700 30 JAMES 03-DEC-81 950 2200
- 75. SQL> select deptno, ename, hiredate, sal, 2 sum(sal) over ( 3 partition by deptno 4 order by hiredate 5 rows between 1 preceding and 1 following) as x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- ---------- 10 CLARK 09-JUN-81 2450 7450 10 KING 17-NOV-81 5000 8750 10 MILLER 23-JAN-82 1300 6300 20 SMITH 17-DEC-80 800 3775 20 JONES 02-APR-81 2975 6775 20 FORD 03-DEC-81 3000 8975 20 SCOTT 09-DEC-82 3000 7100 20 ADAMS 12-JAN-83 1100 4100 30 ALLEN 20-FEB-81 1600 2850 30 WARD 22-FEB-81 1250 5700 30 BLAKE 01-MAY-81 2850 5600 30 TURNER 08-SEP-81 1500 5600 30 MARTIN 28-SEP-81 1250 3700 30 JAMES 03-DEC-81 950 2200
- 76. SQL> select deptno, ename, hiredate, sal, 2 sum(sal) over ( 3 partition by deptno 4 order by hiredate 5 rows between 1 preceding and 1 following) as x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- ---------- 10 CLARK 09-JUN-81 2450 7450 10 KING 17-NOV-81 5000 8750 10 MILLER 23-JAN-82 1300 6300 20 SMITH 17-DEC-80 800 3775 20 JONES 02-APR-81 2975 6775 20 FORD 03-DEC-81 3000 8975 20 SCOTT 09-DEC-82 3000 7100 20 ADAMS 12-JAN-83 1100 4100 30 ALLEN 20-FEB-81 1600 2850 30 WARD 22-FEB-81 1250 5700 30 BLAKE 01-MAY-81 2850 5600 30 TURNER 08-SEP-81 1500 5600 30 MARTIN 28-SEP-81 1250 3700 30 JAMES 03-DEC-81 950 2200
- 77. 77 "I need a 6 month moving average"
- 78. SQL> select deptno, ename, hiredate, sal, 2 avg(sal) over ( 3 partition by deptno 4 order by hiredate 5 range between interval '6' month preceding and current row ) x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- -------- 10 CLARK 09-JUN-81 2450 2450 10 KING 17-NOV-81 5000 3725 10 MILLER 23-JAN-82 1300 3150 20 SMITH 17-DEC-80 800 800 20 JONES 02-APR-81 2975 1888 20 FORD 03-DEC-81 3000 3000 20 SCOTT 09-DEC-82 3000 3000 20 ADAMS 12-JAN-83 1100 2050 30 ALLEN 20-FEB-81 1600 1600 30 WARD 22-FEB-81 1250 1425 30 BLAKE 01-MAY-81 2850 1900 30 TURNER 08-SEP-81 1500 2175 30 MARTIN 28-SEP-81 1250 1867 30 JAMES 03-DEC-81 950 1233
- 79. SQL> select deptno, ename, hiredate, sal, 2 avg(sal) over ( 3 partition by deptno 4 order by hiredate 5 range between interval '6' month preceding and current row ) x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- -------- 10 CLARK 09-JUN-81 2450 2450 10 KING 17-NOV-81 5000 3725 10 MILLER 23-JAN-82 1300 3150 20 SMITH 17-DEC-80 800 800 20 JONES 02-APR-81 2975 1888 20 FORD 03-DEC-81 3000 3000 20 SCOTT 09-DEC-82 3000 3000 20 ADAMS 12-JAN-83 1100 2050 30 ALLEN 20-FEB-81 1600 1600 30 WARD 22-FEB-81 1250 1425 30 BLAKE 01-MAY-81 2850 1900 30 TURNER 08-SEP-81 1500 2175 30 MARTIN 28-SEP-81 1250 1867 30 JAMES 03-DEC-81 950 1233
- 80. SQL> select deptno, ename, hiredate, sal, 2 avg(sal) over ( 3 partition by deptno 4 order by hiredate 5 range between interval '6' month preceding and current row ) x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- -------- 10 CLARK 09-JUN-81 2450 2450 10 KING 17-NOV-81 5000 3725 10 MILLER 23-JAN-82 1300 3150 20 SMITH 17-DEC-80 800 800 20 JONES 02-APR-81 2975 1888 20 FORD 03-DEC-81 3000 3000 20 SCOTT 09-DEC-82 3000 3000 20 ADAMS 12-JAN-83 1100 2050 30 ALLEN 20-FEB-81 1600 1600 30 WARD 22-FEB-81 1250 1425 30 BLAKE 01-MAY-81 2850 1900 30 TURNER 08-SEP-81 1500 2175 30 MARTIN 28-SEP-81 1250 1867 30 JAMES 03-DEC-81 950 1233
- 81. SQL> select deptno, ename, hiredate, sal, 2 avg(sal) over ( 3 partition by deptno 4 order by hiredate 5 range between interval '6' month preceding and current row ) x 6 from emp 7 order by deptno, hiredate; DEPTNO ENAME HIREDATE SAL X ---------- ---------- --------- ---------- -------- 10 CLARK 09-JUN-81 2450 2450 10 KING 17-NOV-81 5000 3725 10 MILLER 23-JAN-82 1300 3150 20 SMITH 17-DEC-80 800 800 20 JONES 02-APR-81 2975 1888 20 FORD 03-DEC-81 3000 3000 20 SCOTT 09-DEC-82 3000 3000 20 ADAMS 12-JAN-83 1100 2050 30 ALLEN 20-FEB-81 1600 1600 30 WARD 22-FEB-81 1250 1425 30 BLAKE 01-MAY-81 2850 1900 30 TURNER 08-SEP-81 1500 2175 30 MARTIN 28-SEP-81 1250 1867 30 JAMES 03-DEC-81 950 1233
- 82. 82 windows can be dynamic
- 83. 83 "I need the sum sales from the previous week day "
- 84. SQL> create or replace 2 function LAST_CLOSE(p_purchase_date date) 3 return number is 4 begin 5 return 6 case to_char(p_purchase_date,'DY') 7 when 'SUN' then 2 -- Fri,Sat 8 when 'MON' then 3 -- Fri,Sat,Sun 9 else 1 10 end; 11 end; 12 / Function created. SQL> select 2 prod_id, cust_id, 3 sum(amount_sold) 4 over ( order by purchase_date 5 range between LAST_CLOSE(purchase_date) preceding) as bus_tot 6 from sales 7 /
- 87. 87 "compare each salary with first person hired across entire organisation and within each department"
- 88. SQL> select deptno, hiredate, empno, ename, sal, 2 first_value(sal) over ( order by hiredate 3 range unbounded preceding ) lo_sal, 4 first_value(sal) over ( partition by deptno 5 order by hiredate 6 range unbounded preceding) lo_dept_sal 7 from emp 8 order by deptno, hiredate; DEPTNO HIREDATE EMPNO ENAME SAL ---------- --------- ---------- ---------- ---------- 10 09-JUN-81 7782 CLARK 2450 10 17-NOV-81 7839 KING 5000 10 23-JAN-82 7934 MILLER 1300 20 17-DEC-80 7369 SMITH 800 20 02-APR-81 7566 JONES 2975 20 03-DEC-81 7902 FORD 3000 20 09-DEC-82 7788 SCOTT 3000 20 12-JAN-83 7876 ADAMS 1100 30 20-FEB-81 7499 ALLEN 1600 30 22-FEB-81 7521 WARD 1250 30 01-MAY-81 7698 BLAKE 2850 30 08-SEP-81 7844 TURNER 1500 30 28-SEP-81 7654 MARTIN 1250 30 03-DEC-81 7900 JAMES 950 LO_SAL ---------- 800 800 800 800 800 800 800 800 800 800 800 800 800 800 LO_DEPT_SAL ----------- 2450 2450 2450 800 800 800 800 800 1600 1600 1600 1600 1600 1600
- 89. 89 beyond first and last
- 90. 90 nth_value
- 91. 91 "compare each salary with second hired person across entire organisation and within each department"
- 92. SQL> select deptno, hiredate, empno, ename, sal, 2 100 * sal / nth_value(sal,2) over ( order by hiredate 3 range unbounded preceding ) sal_pct, 4 100 * sal / nth_value(sal,2) over ( partition by deptno 5 order by hiredate 6 range unbounded preceding) dept_sal_pct 7 from emp 8 order by deptno, hiredate; DEPTNO HIREDATE EMPNO ENAME SAL ---------- --------- ---------- ---------- ---------- 10 09-JUN-81 7782 CLARK 2450 10 17-NOV-81 7839 KING 5000 10 23-JAN-82 7934 MILLER 1300 20 17-DEC-80 7369 SMITH 800 20 02-APR-81 7566 JONES 2975 20 03-DEC-81 7902 FORD 3000 20 09-DEC-82 7788 SCOTT 3000 20 12-JAN-83 7876 ADAMS 1100 30 20-FEB-81 7499 ALLEN 1600 30 22-FEB-81 7521 WARD 1250 30 01-MAY-81 7698 BLAKE 2850 30 08-SEP-81 7844 TURNER 1500 30 28-SEP-81 7654 MARTIN 1250 30 03-DEC-81 7900 JAMES 950 SAL_PCT ------- 153.13 312.50 81.25 185.94 187.50 187.50 68.75 100.00 78.13 178.13 93.75 78.13 59.38 DEPT_SAL_PCT ------------ 100.00 26.00 100.00 100.84 100.84 36.97 100.00 228.00 120.00 100.00 76.00
- 94. SQL> select empno, ename, sal, deptno 2 from emp 3 order by sal; EMPNO ENAME SAL DEPTNO ---------- ---------- ---------- ---------- 7369 SMITH 800 10 7900 JAMES 950 7876 ADAMS 1100 7521 WARD 1250 20 7654 MARTIN 1250 7934 MILLER 1300 7844 TURNER 1500 7499 ALLEN 1600 30 7782 CLARK 2450 7698 BLAKE 2850 7566 JONES 2975 7788 SCOTT 3000 7902 FORD 3000 7839 KING 5000 40
- 95. SQL> select empno, ename, sal, deptno, 2 last_value(deptno IGNORE NULLS) 3 over (order by sal) as last_dept 4 from emp 5 order by sal EMPNO ENAME SAL DEPTNO ---------- ---------- ---------- ---------- 7369 SMITH 800 10 7900 JAMES 950 7876 ADAMS 1100 7521 WARD 1250 20 7654 MARTIN 1250 7934 MILLER 1300 7844 TURNER 1500 7499 ALLEN 1600 30 7782 CLARK 2450 7698 BLAKE 2850 7566 JONES 2975 7788 SCOTT 3000 7902 FORD 3000 7839 KING 5000 40 LAST_DEPT ---------- 10 10 10 20 20 20 20 30 30 30 30 30 30 40
- 96. SQL> select last_dept, count(*) 2 from 3 ( select 4 last_value(deptno ignore nulls) 5 over (order by sal) as last_dept 6 from emp2 7 ) 8 group by last_dept; LAST_DEPT COUNT(*) ---------- ---------- 30 6 20 4 40 1 10 3
- 98. 98 recall
- 99. SQL> select ename, deptno, 2 sum(sal) over ( partition by deptno ) as deptsal 3 from emp 4 order by deptno; ENAME DEPTNO DEPTSAL ---------- ---------- ---------- CLARK 10 8750 KING 10 8750 MILLER 10 8750 JONES 20 10875 FORD 20 10875 ADAMS 20 10875 SMITH 20 10875 SCOTT 20 10875 WARD 30 9400 TURNER 30 9400 ALLEN 30 9400 JAMES 30 9400 BLAKE 30 9400 MARTIN 30 9400 reporting aggregate
- 100. DEPTNO EMPNO ENAME JOB SAL ---------- ---------- ---------- --------- ---------- 10 7782 CLARK MANAGER 2450 10 7839 KING PRESIDENT 5000 10 7934 MILLER CLERK 1300 20 7876 ADAMS CLERK 1100 20 7902 FORD ANALYST 3000 20 7566 JONES MANAGER 2975 20 7788 SCOTT ANALYST 3000 20 7369 SMITH CLERK 800 30 7499 ALLEN SALESMAN 1600 30 7698 BLAKE MANAGER 2850 30 7900 JAMES CLERK 950 30 7654 MARTIN SALESMAN 1250 30 7844 TURNER SALESMAN 1500 30 7521 WARD SALESMAN 1250 RUNNING_TOTAL ------------- 2450 7450 8750 1100 4100 7075 10075 10875 1600 4450 5400 6650 8150 9400 SQL> select deptno, empno, ename, job, sal, 2 sum(sal) OVER ( 3 partition by deptno 4 order by ename) as running_total 5 from emp windowing aggregate
- 101. <function> ( <arg>,<arg>,… ) OVER ( <partition clause> <sorting clause> <windowing clause> ) but I didn't specify one of these ?
- 102. <function> ( <arg>,<arg>,… ) OVER ( <partition clause> <sorting clause> <windowing clause> ) THEN you get one of these automatically ! IF this is an aggregate function ... AND you have included an ORDER BY clause ... range between unbounded preceding and current row
- 103. sum(sal) OVER ( partition by deptno ) function partition clause order by ename
- 104. 104 "OVER"
- 105. 105 lag / lead
- 106. SQL> select 2 empno, ename, hiredate, sal, 3 lag(sal,1) 4 over ( order by hiredate ) prev_hire_sal 5 from emp 6 order by hiredate; EMPNO ENAME HIREDATE SAL ---------- ---------- --------- ---------- 7369 SMITH 17-DEC-80 800 7499 ALLEN 20-FEB-81 1600 7521 WARD 22-FEB-81 1250 7566 JONES 02-APR-81 2975 7698 BLAKE 01-MAY-81 2850 7782 CLARK 09-JUN-81 2450 7844 TURNER 08-SEP-81 1500 7654 MARTIN 28-SEP-81 1250 7839 KING 17-NOV-81 5000 7900 JAMES 03-DEC-81 950 7902 FORD 03-DEC-81 3000 7934 MILLER 23-JAN-82 1300 7788 SCOTT 09-DEC-82 3000 7876 ADAMS 12-JAN-83 1100 PREV_HIRE_SAL -------------- 800 1600 1250 2975 2850 2450 1500 1250 5000 950 3000 1300 3000
- 107. 107 ignore nulls as well
- 108. 108 by default
- 109. SQL> select empno, ename, sal 2 from emp 3 order by empno; EMPNO ENAME SAL ---------- ---------- ---------- 7369 SMITH 800 7499 ALLEN 7521 WARD 7566 JONES 7654 MARTIN 1250 7698 BLAKE 7782 CLARK 7788 SCOTT 7839 KING 5000 7844 TURNER 7876 ADAMS 7900 JAMES 7902 FORD 7934 MILLER
- 110. SQL> select empno, ename, sal, 2 lag(sal,1) 3 over ( order by empno) as prev_sal 4 from emp 5 order by empno; EMPNO ENAME SAL PREV_SAL ---------- ---------- ---------- ---------- 7369 SMITH 800 7499 ALLEN 800 7521 WARD 7566 JONES 7654 MARTIN 1250 7698 BLAKE 1250 7782 CLARK 7788 SCOTT 7839 KING 5000 7844 TURNER 5000 7876 ADAMS 7900 JAMES 7902 FORD 7934 MILLER
- 111. SQL> select empno, ename, sal, 2 lag(sal,1) ignore nulls 3 over ( order by empno) as prev_sal 4 from emp 5 order by empno; EMPNO ENAME SAL ---------- ---------- ---------- 7369 SMITH 800 7499 ALLEN 7521 WARD 7566 JONES 7654 MARTIN 1250 7698 BLAKE 7782 CLARK 7788 SCOTT 7839 KING 5000 7844 TURNER 7876 ADAMS 7900 JAMES 7902 FORD 7934 MILLER PREV_SAL ---------- 800 800 800 800 1250 1250 1250 1250 5000 5000 5000 5000 5000
- 112. 112 and then....
- 113. 113 “imagination is more important than knowledge” - Albert Einstein
- 117. SQL> select * from BAD_EMP; EMPNO ENAME JOB MGR HIREDATE SAL ---------- ---------- --------- ---------- --------- ---------- 7788 SCOTT ANALYST 7566 09-DEC-82 3000 7499 ALLEN SALESMAN 7698 20-FEB-81 1600 7369 SMITH CLERK 7902 17-DEC-80 800 7839 KING PRESIDENT 17-NOV-81 5000 7566 JONES MANAGER 7839 02-APR-81 2975 7876 ADAMS CLERK 7788 12-JAN-83 1100 7844 TURNER SALESMAN 7698 08-SEP-81 1500 7499 ALLEN SALESMAN 7698 20-FEB-81 1600 7369 SMITH CLERK 7902 17-DEC-80 800 7902 FORD ANALYST 7566 03-DEC-81 3000 7900 JAMES CLERK 7698 03-DEC-81 950 7521 WARD SALESMAN 7698 22-FEB-81 1250 7654 MARTIN SALESMAN 7698 28-SEP-81 1250 7934 MILLER CLERK 7782 23-JAN-82 1300 7698 BLAKE MANAGER 7839 01-MAY-81 2850 7782 CLARK MANAGER 7839 09-JUN-81 2450
- 118. SQL> select * from BAD_EMP; EMPNO ENAME JOB MGR HIREDATE SAL ---------- ---------- --------- ---------- --------- ---------- 7788 SCOTT ANALYST 7566 09-DEC-82 3000 7499 ALLEN SALESMAN 7698 20-FEB-81 1600 7369 SMITH CLERK 7902 17-DEC-80 800 7839 KING PRESIDENT 17-NOV-81 5000 7566 JONES MANAGER 7839 02-APR-81 2975 7876 ADAMS CLERK 7788 12-JAN-83 1100 7844 TURNER SALESMAN 7698 08-SEP-81 1500 7499 ALLEN SALESMAN 7698 20-FEB-81 1600 7369 SMITH CLERK 7902 17-DEC-80 800 7902 FORD ANALYST 7566 03-DEC-81 3000 7900 JAMES CLERK 7698 03-DEC-81 950 7521 WARD SALESMAN 7698 22-FEB-81 1250 7654 MARTIN SALESMAN 7698 28-SEP-81 1250 7934 MILLER CLERK 7782 23-JAN-82 1300 7698 BLAKE MANAGER 7839 01-MAY-81 2850 7782 CLARK MANAGER 7839 09-JUN-81 2450
- 119. SQL> select empno, rowid, 2 row_number() over 3 ( partition by empno order by rowid ) as r 4 from BAD_EMP 5 / EMPNO ROWID R ---------- ------------------ ---------- 7369 AAARXwAAEAAATlMAAA 1 7369 AAARXwAAEAAATlOAAA 2 7499 AAARXwAAEAAATlMAAB 1 7499 AAARXwAAEAAATlOAAB 2 7521 AAARXwAAEAAATlMAAC 1 7566 AAARXwAAEAAATlMAAD 1 7654 AAARXwAAEAAATlMAAE 1 7698 AAARXwAAEAAATlMAAF 1 7782 AAARXwAAEAAATlMAAG 1 7788 AAARXwAAEAAATlMAAH 1 7839 AAARXwAAEAAATlMAAI 1 7844 AAARXwAAEAAATlMAAJ 1 7876 AAARXwAAEAAATlMAAK 1 7900 AAARXwAAEAAATlMAAL 1 7902 AAARXwAAEAAATlMAAM 1 7934 AAARXwAAEAAATlMAAN 1
- 120. SQL> delete from BAD_EMP 2 where ROWID in 3 ( select rowid 4 from 5 ( select rowid, 6 row_number() over 7 ( partition by empno 8 order by rowid) as r 9 from BAD_EMP 10 ) 11 where r > 1 12 ) 13 / 2 rows deleted.
- 121. 121 “5 highest salaries from each department”
- 122. SQL> select deptno, salary 2 from 3 ( select 4 deptno, 5 salary, 6 rank() over ( 7 partition by deptno 8 order by salary) top_5 9 from EMPLOYEES 10 ) 11 where top_5 <= 5
- 123. 123 “mind the gap”
- 124. 124 SQL> select X from T; X ---------- 2 3 4 7 8 12 13 15 16 17 19 20 2-4 7-8 12-13 15-17 19-20
- 125. 125 SQL> select 2 x, 3 lag(x) over ( order by x) prev 4 from T ; X PREV ---------- ---------- 2 3 2 4 3 7 4 8 7 12 8 13 12 15 13 16 15 17 16 19 17 20 19
- 126. 126 SQL> select 2 x, 3 lag(x) over ( order by x) prev 4 from T ; X PREV ---------- ---------- 2 3 2 4 3 7 4 8 7 12 8 13 12 15 13 16 15 17 16 19 17 20 19
- 127. 127 SQL> select 2 x, 3 lag(x) over ( order by x) prev 4 from T ; X PREV ---------- ---------- 2 3 2 4 3 7 4 8 7 12 8 13 12 15 13 16 15 17 16 19 17 20 19
- 128. 128 SQL> select 2 x, 3 lag(x) over ( order by x) prev 4 from T ; X PREV ---------- ---------- 2 3 2 4 3 7 4 8 7 12 8 13 12 15 13 16 15 17 16 19 17 20 19
- 129. 129 SQL> select 2 x, 3 lag(x) over ( order by x) prev 4 from T ; X PREV ---------- ---------- 2 3 2 4 3 7 4 8 7 12 8 13 12 15 13 16 15 17 16 19 17 20 19
- 130. 130 SQL> select 2 x, 3 case 4 when nvl(lag(x) over (order by x),x) != x-1 5 then x end loval 6 from t; X LOVAL ---------- ---------- 2 2 3 4 7 7 8 12 12 13 15 15 16 17 19 19 20
- 131. 131 SQL> select 2 x, 3 case 4 when nvl(lag(x) over (order by x),x) != x-1 5 then x end loval 6 from t; X LOVAL ---------- ---------- 2 2 3 4 7 7 8 12 12 13 15 15 16 17 19 19 20
- 132. 132 SQL> select 2 x, 3 case 4 when nvl(lag(x) over (order by x),x) != x-1 5 then x end loval 6 from t; X LOVAL ---------- ---------- 2 2 3 4 7 7 8 12 12 13 15 15 16 17 19 19 20
- 133. 133 SQL> select 2 x, 3 case 4 when nvl(lag(x) over (order by x),x) != x-1 5 then x end loval 6 from t; X LOVAL ---------- ---------- 2 2 3 4 7 7 8 12 12 13 15 15 16 17 19 19 20
- 134. 134 SQL> select x, max(loval) over (order by x) loval 2 from ( 3 select x, 4 case 5 when nvl(lag(x) over (order by x),x) != x-1 6 then x end loval 7 from t ); X LOVAL ---------- ---------- 2 2 3 2 4 2 7 7 8 7 12 12 13 12 15 15 16 15 17 15 19 19 20 19
- 135. 135 SQL> select x, max(loval) over (order by x) loval 2 from ( 3 select x, 4 case 5 when nvl(lag(x) over (order by x),x) != x-1 6 then x end loval 7 from t ); X LOVAL ---------- ---------- 2 2 3 2 4 2 7 7 8 7 12 12 13 12 15 15 16 15 17 15 19 19 20 19
- 136. 136 SQL> select min(x)||’-’||max(x) ranges from ( 2 select x,max(loval) over (order by x) loval 3 from ( 4 select x, 5 case 6 when nvl(lag(x) over (order by x),x) != x-1 7 then x end loval 8 from t)) 9 group by loval; RANGES -------------------- 2-4 7-8 12-13 15-17 19-20
- 137. 137 or ...
- 138. 138 SQL> select 2 min(x)||'-'|| 3 case when min (x) = max (x) 4 then min(x) 5 else max(x) 6 end rng 7 from 8 (select X 9 , row_number() over (order by X) rn 10 from t 11 ) 12 group by x - rn 13 order by min(x); RNG ---------------- 2-4 7-8 12-13 15-17 19-20
- 140. sql_string = "select * from ACCOUNTS where ACCT_NO in ( :bindvar )" EXEC SQL PREPARE sql_string; ORA-01722: invalid number 123,456,789
- 141. SQL> exec :acct = '123,456,789' SQL> select substr(:acct, 2 loc+1,nvl( 3 lead(loc) over ( order by loc ) – loc-1, 4 length(:acct)-loc) 5 ) list_as_rows 6 from ( 7 select distinct (instr(:acct,',',1,level)) loc 8 from dual 9 connect by level < length(:acct)- 10 length(replace(:acct,','))+1 11 ); LIST_AS_ROWS -------------------- 123 456 789
- 142. SQL> with MY_LIST as select substr(:acct, 2 loc+1,nvl( 3 lead(loc) over ( order by loc ) – loc-1, 4 length(:acct)-loc) 5 ) val 6 from ( 7 select distinct (instr(:acct,',',1,level)) loc 8 from dual 9 connect by level < length(:acct)- 10 length(replace(:acct,','))+1 11 ) 12 select * 13 from ACCOUNTS 14 where ACCT_NO in ( select val from MY_LIST)
- 144. 144 opposite
- 145. 145 listagg
- 147. 147 SQL> select deptno, ename 2 from emp 3 order by 1,2; DEPTNO ENAME ---------- ---------- 10 CLARK 10 KING 10 MILLER 20 ADAMS 20 FORD 20 JONES 20 SCOTT 20 SMITH 30 ALLEN 30 BLAKE 30 JAMES 30 MARTIN 30 TURNER 30 WARD
- 148. 148 DEPTNO MEMBERS ---------- ------------------------------------- 10 CLARK,KING,MILLER 20 SMITH,JONES,SCOTT,ADAMS,FORD 30 ALLEN,WARD,MARTIN,BLAKE,TURNER,JAMES
- 149. 149 what we used to do
- 150. SQL> select deptno , rtrim(ename,',') enames 2 from ( select deptno,ename,rn 3 from emp 4 model 5 partition by (deptno) 6 dimension by ( 7 row_number() over 8 (partition by deptno order by ename) rn 9 ) 10 measures (cast(ename as varchar2(40)) ename) 11 rules 12 ( ename[any] 13 order by rn desc = ename[cv()]||','||ename[cv()+1]) 14 ) 15 where rn = 1 16 order by deptno; DEPTNO ENAMES ---------- ---------------------------------------- 10 CLARK,KING,MILLER 20 ADAMS,FORD,JONES,SCOTT,SMITH 30 ALLEN,BLAKE,JAMES,MARTIN,TURNER,WARD - Rob Van Wijk
- 151. SQL> select deptno, 2 substr(max(sys_connect_by_path(ename, ',')), 2) members 3 from (select deptno, ename, 4 row_number () 5 over (partition by deptno order by empno) rn 6 from emp) 7 start with rn = 1 8 connect by prior rn = rn - 1 9 and prior deptno = deptno 10 group by deptno 11 / DEPTNO MEMBERS ---------- --------------------------------------------------------- 30 ALLEN,WARD,MARTIN,BLAKE,TURNER,JAMES 20 SMITH,JONES,SCOTT,ADAMS,FORD 10 CLARK,KING,MILLER - Anon
- 152. SQL> select deptno, 2 xmltransform 3 ( sys_xmlagg 4 ( sys_xmlgen(ename) 5 ), 6 xmltype 7 ( 8 '<?xml version="1.0"?><xsl:stylesheet version="1.0" 9 xmlns:xsl="http://www.w3.org/1999/XSL/Transform"> 10 <xsl:template match="/"> 11 <xsl:for-each select="/ROWSET/ENAME"> 12 <xsl:value-of select="text()"/>;</xsl:for-each> 13 </xsl:template> 14 </xsl:stylesheet>' 15 ) 16 ).getstringval() members 17 from emp 18 group by deptno; DEPTNO MEMBERS ---------- -------------------------------------------------------- 10 CLARK;MILLER;KING; 20 SMITH;FORD;ADAMS;SCOTT;JONES; 30 ALLEN;JAMES;TURNER;BLAKE;MARTIN;WARD; - Laurent Schneider
- 153. SQL> create or replace type string_agg_type as object 2 ( 3 total varchar2(4000), 4 5 static function 6 ODCIAggregateInitialize(sctx IN OUT string_agg_type ) 7 return number, 8 9 member function 10 ODCIAggregateIterate(self IN OUT string_agg_type , 11 value IN varchar2 ) 12 return number, 13 14 member function 15 ODCIAggregateTerminate(self IN string_agg_type, 16 returnValue OUT varchar2, 17 flags IN number) 18 return number, 19 20 member function 21 ODCIAggregateMerge(self IN OUT string_agg_type, 22 ctx2 IN string_agg_type) 23 return number 24 ); 25 /
- 154. 154 hard
- 155. 155 SQL> select deptno, 2 listagg( ename, ',') 3 within group (order by empno) members 4 from emp 5 group by deptno; DEPTNO MEMBERS ---------- ----------------------------------------- 10 CLARK,KING,MILLER 20 SMITH,JONES,SCOTT,ADAMS,FORD 30 ALLEN,WARD,MARTIN,BLAKE,TURNER,JAMES
- 156. 156 other goodies
- 157. 157 KEEP extension
- 158. “Show me lowest salary for each department...” SQL> select deptno, min(sal) 2 from emp 3 group by deptno; “...and I need to know who has that lowest salary as well” SQL> select deptno, empno, min(sal) 2 from emp 3 group by deptno; ORA-00979: not a GROUP BY expression
- 159. 159 SQL> select deptno, min(sal), min(empno) 2 KEEP ( dense_rank FIRST order by sal) empno 3 from emp 4 group by deptno 5 / DEPTNO MIN(SAL) EMPNO ---------- ---------- ---------- 10 1300 7934 20 800 7369 30 950 7900 Emp 7934 has the lowest salary in dept 10
- 161. 161 recall
- 162. 162 cume_dist
- 163. 163 SQL> select ename, sal, 2 100*cume_dist() over ( order by sal ) as pct 3 from emp 4 order by sal; ENAME SAL PCT ---------- ---------- ------- SMITH 800 7.14 JAMES 950 14.29 ADAMS 1100 21.43 WARD 1250 35.71 MARTIN 1250 35.71 MILLER 1300 42.86 TURNER 1500 50.00 ALLEN 1600 57.14 CLARK 2450 64.29 BLAKE 2850 71.43 JONES 2975 78.57 FORD 3000 92.86 SCOTT 3000 92.86 KING 5000 100.00
- 164. 164 "what is the 60th percentile salary in the company?"
- 165. 165 SQL> select ename, sal, 2 100*cume_dist() over ( order by sal ) as pct 3 from emp 4 order by sal; ENAME SAL PCT ---------- ---------- ------- SMITH 800 7.14 JAMES 950 14.29 ADAMS 1100 21.43 WARD 1250 35.71 MARTIN 1250 35.71 MILLER 1300 42.86 TURNER 1500 50.00 ALLEN 1600 57.14 CLARK 2450 64.29 BLAKE 2850 71.43 JONES 2975 78.57 FORD 3000 92.86 SCOTT 3000 92.86 KING 5000 100.00
- 166. 166 SQL> select ename, sal, 2 100*cume_dist() over ( order by sal ) as pct 3 from emp 4 order by sal; ENAME SAL PCT ---------- ---------- ------- SMITH 800 7.14 JAMES 950 14.29 ADAMS 1100 21.43 WARD 1250 35.71 MARTIN 1250 35.71 MILLER 1300 42.86 TURNER 1500 50.00 ALLEN 1600 57.14 CLARK 2450 64.29 BLAKE 2850 71.43 JONES 2975 78.57 FORD 3000 92.86 SCOTT 3000 92.86 KING 5000 100.00
- 167. 167 SQL> select 2 percentile_disc(0.6) 3 within group 4 (order by sal) as dicrete_pct, 5 percentile_cont(0.6) 6 within group 7 (order by sal) as continuous_pct 8 from emp; DICRETE_PCT CONTINUOUS_PCT ----------- -------------- 2450 2280
- 169. 169 "median salary for each department?"
- 170. 170 what we used to do
- 171. SQL> select deptno, avg(distinct sal) median 2 from 3 (select cp1.deptno, cp1.sal 4 from emp cp1, emp cp2 5 where cp1.deptno = cp2.deptno 6 group by cp1.deptno, cp1.sal 7 having sum(decode(cp1.sal, cp2.sal, 1, 0)) >= 8 abs(sum(sign(cp1.sal - cp2.sal)))) 9 group by deptno 10 / DEPTNO MEDIAN ---------- ---------- 10 3725 20 2975 30 1375 40 1300 wth?
- 172. SQL> select 2 deptno, 3 percentile_cont(0.5) 4 within group (order by sal) as median 5 from emp 6 group by deptno; SQL> select 2 deptno, 3 median(sal) 4 from emp 5 group by deptno; DEPTNO MEDIAN(SAL) ---------- ----------- 10 3725 20 2975 30 1375 40 1300
- 174. 174 "if I was paid $3000, where would I rank in each department?"
- 175. 175 SQL> select 2 deptno, 3 rank(3000) within group 4 ( order by sal ) as ranking 5 from emp 6 group by deptno; DEPTNO RANKING ---------- ---------- 10 2 20 4 30 7 40 2
- 176. 176 ratio_to_report
- 177. 177 "salary percentage breakdown across employees"
- 178. SQL> select 2 empno, 3 ename, 4 sal, 5 100*ratio_to_report(sal) over () as pct 6 from emp; EMPNO ENAME SAL PCT ---------- ---------- ---------- ------- 7521 WARD 1250 4.69 7566 JONES 2975 11.17 7654 MARTIN 1250 4.69 7698 BLAKE 2850 10.70 7782 CLARK 2450 9.20 7788 SCOTT 3000 11.27 7839 KING 5000 18.78 7844 TURNER 1500 5.63 7876 ADAMS 1100 4.13 7900 JAMES 950 3.57 7902 FORD 3000 11.27 7934 MILLER 1300 4.88
- 179. 179 3 more things ... not really analytic ?
- 181. 181 SQL> select * 2 from timeslots; HR -- 8 9 10 11 12 13 14 15 16 SQL> select * 2 from bookings; HR ROOM WHO ------- ---------- ------- 8 Room2 PETE 9 Room1 JOHN 11 Room1 MIKE 14 Room2 JILL 15 Room2 JANE 16 Room1 SAM
- 182. 182 bookings by hour conventional outer join
- 183. 183 SQL> SELECT hrs.hr, t1.room, t1.who 2 from timeslots hrs 3 left outer join bookings t1 4 on hrs.hr = t1.hr 5 order by 1 HR ROOM WHO ------- ---------- ---------- 8 Room2 PETE 9 Room1 JOHN 10 11 Room1 MIKE 12 13 14 Room2 JILL 15 Room2 JANE 16 Room1 SAM
- 184. 184 bookings by hour per room
- 185. 185 HR ROOM WHO ------- ---------- ---------- 8 Room2 PETE 9 10 11 12 13 14 Room2 JILL 15 Room2 JANE 16 HR ROOM WHO ------- ---------- ---------- 8 9 Room1 JOHN 10 11 Room1 MIKE 12 13 14 15 16 Room1 SAM
- 186. 186 SQL> select * 2 from timeslots; HR -- 8 9 10 11 12 13 14 15 16 x "Room 1" x "Room 2" ... x "Room n"
- 188. 188 SQL> SELECT hrs.hr, t1.room, t1.who 2 FROM bookings t1 3 PARTITION BY (t1.room) 4 RIGHT OUTER JOIN timeslots ON (hrs.hr = t1.hr) 5 order by 1,2 HR ROOM WHO --------- ---------- ---------- 8 Room1 9 Room1 JOHN 10 Room1 11 Room1 MIKE 12 Room1 13 Room1 14 Room1 15 Room1 16 Room1 SAM 8 Room2 PETE 9 Room2 10 Room2 11 Room2 12 Room2 13 Room2 14 Room2 JILL 15 Room2 JANE 16 Room2
- 189. 189 width_bucket
- 190. SQL> select 2 object_id, 3 width_bucket(object_id, 4 1000, 5 90000, 6 10) bucket 7 from dba_objects OBJECT_ID BUCKET ---------- ---------- 913 0 ... 3231 1 ... 5858 1 ... 14920 2 ... 42421 5 ... 91635 11 < min > max 1 .. buckets
- 192. 192 12c+
- 193. 193 SQL> select employee_id, last_name 2 from employees 3 order by employee_id 4 fetch first 5 rows only;
- 195. 196 things to note
- 196. 197 cannot be a predicate
- 197. SQL> select ename, deptno, sal 2 from emp 3 where 4 sum(sal) over 5 ( partition by deptno) > 10; sum(sal) over * ERROR at line 4: ORA-00934: group function is not allowed here
- 198. 199 inline view
- 199. 200 SQL> select ename, deptno, sal 2 from ( 3 select ename, deptno, sal, 4 sum(sal) over 5 ( partition by deptno) as deptsal 6 from emp 7 ) 8 where deptsal > 10;
- 201. 202 create view RANKED_ACCOUNTS as select account_num, customer_name, acct_type_code, rank() over ( order by gross_sales ) as seq from ACCOUNTS; indexed column
- 202. 203 SQL> select * from RANKED_ACCOUNTS 2 where ACCOUNT_NUM = 12345 ------------------------------------------------ | Id | Operation | Name | ------------------------------------------------ | 0 | SELECT STATEMENT | | | 1 | VIEW | RANKED_ACCOUNTS | | 2 | WINDOW SORT | | | 3 | TABLE ACCESS FULL| ACCOUNTS | ------------------------------------------------
- 203. 204 a more holistic view
- 204. recall
- 205. SQL> select deptno, 2 listagg( ename, ',') 3 within group (order by empno) 4 from emp 5 group by deptno;
- 206. 207 still challenges
- 207. 208 real example
- 209. 210 "Find 10 consecutive deposits in a 24 hour period, then a withdrawal within three days of the last deposit, at a different branch"
- 210. ACCT TSTAMP WTHD_TSTAMP T AMT ---------- ------------------ ------------------ - ---------- 54261 25/01/13 17:20:55 D 100 54261 25/01/13 17:56:58 D 165 54261 26/01/13 11:24:14 D 30 54261 26/01/13 11:47:53 D 45 54261 26/01/13 12:59:38 D 100 54261 26/01/13 13:26:04 D 80 54261 26/01/13 14:41:09 D 50 54261 26/01/13 14:53:12 D 50 54261 26/01/13 15:15:05 D 50 54261 26/01/13 15:51:17 D 50 54261 26/01/13 16:15:02 D 120 54261 26/01/13 16:36:51 D 100 54261 26/01/13 16:55:09 D 100 54261 26/01/13 18:07:17 26/01/13 18:07:17 W -500
- 211. 212 hard...
- 213. 214 12c
- 214. 215 pattern matching
- 215. SQL> select acct, tstamp, wthd_tstamp, txn_type, amt 2 from account_txns 3 MATCH_RECOGNIZE 4 ( 5 partition by acct 6 order by tstamp 7 measures 8 dep.tstamp dep_tstamp, 9 wthd.tstamp wthd_tstamp 10 11 all rows per match 12 pattern ( dep{10,} wthd ) 13 define 14 dep as 15 txn_type = 'D', 16 wthd as 17 txn_type = 'W' 18 and last(dep.tstamp)-first(dep.tstamp) < interval '1' day 19 and wthd.tstamp - last(dep.tstamp) < interval '3' day 20 and wthd.loc_id != last(dep.loc_id) 21 )
- 216. SQL> select acct, tstamp, wthd_tstamp, txn_type, amt 2 from account_txns 3 MATCH_RECOGNIZE 4 ( 5 partition by acct 6 order by tstamp 7 measures 8 dep.tstamp dep_tstamp, 9 wthd.tstamp wthd_tstamp 10 11 all rows per match 12 pattern ( dep{10,} wthd ) 13 define 14 dep as 15 txn_type = 'D', 16 wthd as 17 txn_type = 'W' 18 and last(dep.tstamp)-first(dep.tstamp) < interval '1' day 19 and wthd.tstamp - last(dep.tstamp) < interval '3' day 20 and wthd.loc_id != last(dep.loc_id) 21 )
- 217. SQL> select acct, tstamp, wthd_tstamp, txn_type, amt 2 from account_txns 3 MATCH_RECOGNIZE 4 ( 5 partition by acct 6 order by tstamp 7 measures 8 dep.tstamp dep_tstamp, 9 wthd.tstamp wthd_tstamp 10 11 all rows per match 12 pattern ( dep{10,} wthd ) 13 define 14 dep as 15 txn_type = 'D', 16 wthd as 17 txn_type = 'W' 18 and last(dep.tstamp)-first(dep.tstamp) < interval '1' day 19 and wthd.tstamp - last(dep.tstamp) < interval '3' day 20 and wthd.loc_id != last(dep.loc_id) 21 )
- 218. 219 wrap up
- 219. 220 analytics
- 220. 221 super cool
- 221. 222 super powerful
- 222. 223 less code
- 223. 224 easier to read code
- 224. 225 Thank you youtube bit.ly/youtube-connor blog bit.ly/blog-connor twitter bit.ly/twitter-connor