cari

Rumah >pangkalan data >tutorial mysql >Oracle分析函数/聚合函数使用总结 .

Oracle分析函数/聚合函数使用总结 .

WBOY
WBOYasal
2016-06-07 15:32:231185semak imbas

总结: group by rollup(field1,field2); group by cube(field1,field2); group by grouping sets(field1,field2); 生成测试脚本： [c-sharp:nogutter:collapse:showcolumns] view plaincopyprint? ·········10········20········30

总结:

　group by rollup(field1,field2);　

   group by cube(field1,field2);　

   group by grouping sets(field1,field2);

 

 

生成测试脚本：

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. SQL> CREATE TABLE Bill  
  2.   2  (Bill_Month VARCHAR2(6),    
  3.   3  Area_Code INTEGER,  
  4.   4  Net_Type CHAR(1),  
  5.   5   Local_Fare NUMBER(10,2))  

插入测试数据：

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5761,'J',5667089.85 );  
  2. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5762,'G',6315075.96 );  
  3. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5762,'J',6328716.15 );  
  4. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5763,'G',8861742.59 );  
  5. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5763,'J',7788036.32 );  
  6. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5764,'G',6028670.45 );  
  7. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5764,'J',6459121.49 );  
  8. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5765,'G',13156065.77);  
  9. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200405',5765,'J',11901671.70);  
  10. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5761,'G',7614587.96 );  
  11. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5761,'J',5704343.05 );  
  12. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5762,'G',6556992.60 );  
  13. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5762,'J',6238068.05 );  
  14. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5763,'G',9130055.46 );  
  15. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5763,'J',7990460.25 );  
  16. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5764,'G',6387706.01 );  
  17. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5764,'J',6907481.66 );  
  18. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5765,'G',13562968.81);  
  19. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200406',5765,'J',12495492.50);  
  20. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5761,'G',7987050.65 );  
  21. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5761,'J',5723215.28 );  
  22. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5762,'G',6833096.68 );  
  23. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5762,'J',6391201.44 );  
  24. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5763,'G',9410815.91 );  
  25. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5763,'J',8076677.41 );  
  26. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5764,'G',6456433.23 );  
  27. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5764,'J',6987660.53 );  
  28. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5765,'G',14000101.20);  
  29. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200407',5765,'J',12301780.20);  
  30. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5761,'G',8085170.84 );  
  31. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5761,'J',6050611.37 );  
  32. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5762,'G',6854584.22 );  
  33. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5762,'J',6521884.50 );  
  34. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5763,'G',9468707.65 );  
  35. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5763,'J',8460049.43 );  
  36. insert into Bill (Bill_Month,Area_Code,Net_Type,Local_Fare) values('200408',5764,'G',6587559.23 );  

 

1.ROLLUP

ROLLUP是对group by的扩展，因此，它只能出现在group by子句中，依赖于分组的列，对每个分组会生成汇总数据，如下：
SELECT ….
FROM ….
GROUP BY ROLLUP(C1,C2,C3….C(n-1),C(n));
总共会进行n+1个分组,那么实际上有n+1个group by的union all结果。
第1个分组：全分组。C1,C2,C3….C(n-1),C(n)
第2个分组:C1,C2,C3….C(n-1);//这个分组实际上就是对前面前n-1列分组的小计.
----然后逐渐递减分组列
第n个分组:C1。对上一个分组的小计。
第n+1个分组。不分组全量汇总,相当于合计,也是对group by C1的小计,相当于group by null。

 

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. SELECT NVL(TO_CHAR(AREA_CODE), '总计') AREA_CODE,  
  2.        SUM(LOCAL_FARE) LOCAL_FARE  
  3. FROM   BILL  
  4. GROUP  BY ROLLUP(AREA_CODE)  

--result

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1.     AREA_CODE   LOCAL_FARE  
  2. 5764    45814632.6  
  3. 5761    54225413.04  
  4. 5762    52039619.6  
  5. 5763    69186545.02  
  6. 5765    77418080.18  
  7. 合计  298684290.44  

 

2.cube

CUBE(交叉列表)也是对group by运算的一种扩展，它比rollup扩展更加精细，组合类型更多，rollup是按组合的列从右到左递减分组计算，而CUBE则是对所有可能的组合情况进行分组，这样分组的情况更多，覆盖所有的可能分组，并计算所有可能的分组的小计。比如：
CUBE(C1,C2,C3……C(N))对N个列进行CUBE分组，那么可能的分组情况有：
不分组：C(n,0)
取一列分组:C(n,1)
-----
取N列分组，全分组：C(n,n)
那么运用数学上的组合公式，得出所有所有可能的组合方式有:C(n,0)+C(n,1)+….+C(n,n)=2^n种。
我们以前面的rollup组合列为例子：rollup(name,month)是计算按区域名和月份分组以及每个区域的所有月份的小计以及总计。但是使用 cube(name,month)则有4种分组，比rollup多一个每个月的所有区域的小计。下面比较一下这两种分组方式：

 

分组公式        描述
rollup(name,month)        分组情况有：
group by name,month
group by name,null  //每个区域所有月份小计
group by null,null  //合计
cube(name,month)        分组情况有:
group by null,null  //总计
group by null,month //每个月份的所有区域小计
group by name,null //每个区域的所有月份小计
group by name,month
       
CUBE使用方式：
和rollup一样，是
select …
from …
group by cube(分组列列表)

 

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. SELECT (NVL(BILL_MONTH, '月份')) BILL_MONTH,  
  2.        (TO_CHAR(AREA_CODE)) AREA_CODE,  
  3.        SUM(LOCAL_FARE) LOCAL_FARE  
  4. FROM   BILL  
  5. GROUP  BY CUBE(AREA_CODE, BILL_MONTH)  
  6. ORDER  BY BILL_MONTH, AREA_CODE  

--result

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. BILL_MONTH  AREA_CODE   LOCAL_FARE  
  2. 200405  5761    13060433.89  
  3. 200405  5762    12643792.11  
  4. 200405  5763    16649778.91  
  5. 200405  5764    12487791.94  
  6. 200405  5765    25057737.47  
  7. 200405      79899534.32  
  8. 200406  5761    13318931.01  
  9. 200406  5762    12795060.65  
  10. 200406  5763    17120515.71  
  11. 200406  5764    13295187.67  
  12. 200406  5765    26058461.31  
  13. 200406      82588156.35  
  14. 200407  5761    13710265.93  
  15. 200407  5762    13224298.12  
  16. 200407  5763    17487493.32  
  17. 200407  5764    13444093.76  
  18. 200407  5765    26301881.4  
  19. 200407      84168032.53  
  20. 200408  5761    14135782.21  
  21. 200408  5762    13376468.72  
  22. 200408  5763    17928757.08  
  23. 200408  5764    6587559.23  
  24. 200408      52028567.24  
  25. 月份  5761    54225413.04  
  26. 月份  5762    52039619.6  
  27. 月份  5763    69186545.02  
  28. 月份  5764    45814632.6  
  29. 月份  5765    77418080.18  
  30. 月份      298684290.44  

 

扩展一下，GROUPING函数

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. SELECT DECODE(GROUPING(AREA_CODE), 1, 'all area', TO_CHAR(AREA_CODE)) AREA_CODE,  
  2.        DECODE(GROUPING(BILL_MONTH), 1, 'all month', BILL_MONTH) BILL_MONTH,  
  3.        SUM(LOCAL_FARE) LOCAL_FARE  
  4. FROM   bill  
  5. GROUP  BY CUBE(AREA_CODE, BILL_MONTH)  
  6. ORDER  BY AREA_CODE, BILL_MONTH NULLS LAST  

--Result

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. 5761    200405  13060433.89  
  2. 5761    200406  13318931.01  
  3. 5761    200407  13710265.93  
  4. 5761    200408  14135782.21  
  5. 5761    all month   54225413.04  
  6. 5762    200405  12643792.11  
  7. 5762    200406  12795060.65  
  8. 5762    200407  13224298.12  
  9. 5762    200408  13376468.72  
  10. 5762    all month   52039619.6  
  11. 5763    200405  16649778.91  
  12. 5763    200406  17120515.71  
  13. 5763    200407  17487493.32  
  14. 5763    200408  17928757.08  
  15. 5763    all month   69186545.02  
  16. 5764    200405  12487791.94  
  17. 5764    200406  13295187.67  
  18. 5764    200407  13444093.76  
  19. 5764    200408  6587559.23  
  20. 5764    all month   45814632.6  
  21. 5765    200405  25057737.47  
  22. 5765    200406  26058461.31  
  23. 5765    200407  26301881.4  
  24. 5765    all month   77418080.18  
  25. all area    200405  79899534.32  
  26. all area    200406  82588156.35  
  27. all area    200407  84168032.53  
  28. all area    200408  52028567.24  
  29. all area    all month   298684290.44  

 

以上我们已经掌握了rollup,cube分组统计的知识。但是rollup和cube的分组统计包含了常规group by的统计明细以及对相关列的小计和合计值。如果我们需要的只是按每个分组列小计呢？oracle提供了grouping sets操作，对group by的另一个扩展，专门对分组列分别进行小计计算，不包括合计。使用方式和rollup和cube一样，都是放在group by中。如：
grouping sets(C1,C2….Cn)则分组方式有n种,等于列的数目。
group by c1,null,null…..null。
group by null,c2,null….null。
….
group by null,null……..Cn。
无group by null,null….null，也就是说没有合计行。
注意：grouping sets的统计结果和列的顺序无关。

有时我们只需要月、地区统计结果：

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. SELECT DECODE(GROUPING(AREA_CODE), 1, 'all area', TO_CHAR(AREA_CODE)) AREA_CODE,  
  2.        DECODE(GROUPING(BILL_MONTH), 1, 'all month', BILL_MONTH) BILL_MONTH,  
  3.        SUM(LOCAL_FARE) LOCAL_FARE  
  4. FROM   BILL  
  5. GROUP  BY GROUPING SETS(AREA_CODE, BILL_MONTH);  

--result

[c-sharp:nogutter:collapse:showcolumns] view plaincopyprint?

·········10········20········30········40········50········60········70········80········90········100·······110·······120·······130·······140·······150

  1. AREA_CODE   BILL_MONTH  LOCAL_FARE  
  2. 5764    all month   45814632.6  
  3. 5761    all month   54225413.04  
  4. 5762    all month   52039619.6  
  5. 5763    all month   69186545.02  
  6. 5765    all month   77418080.18  
  7. all area    200405  79899534.32  
  8. all area    200406  82588156.35  
  9. all area    200407  84168032.53  
  10. all area    200408  52028567.24  

 

 

3.ROWS

rows是物理行，就是按行的位置，根据位置计算窗口范围

sql query

[c-sharp:collapse] view plaincopyprint?

  1. select bill_month,area_code,net_type,local_fare,sum(local_fare) over(order by local_fare  rows between current row and 1 following) sum_fare  
  2. from bill  
  3. /  
  4. BILL_M  AREA_CODE N LOCAL_FARE   SUM_FARE  
  5. ------ ---------- - ---------- ----------  
  6. 200405       5761 J 5667089.85 11371432.9  
  7. 200406       5761 J 5704343.05 11733013.5  
  8. 200405       5764 G 6028670.45 12266738.5  
  9. 200406       5762 J 6238068.05 12625774.1  
  10. 200406       5764 G 6387706.01 12846827.5  
  11. 200405       5764 J 6459121.49 13016114.1  
  12. 200406       5762 G  6556992.6 13464474.3  
  13. 200406       5764 J 6907481.66 14522069.6  
  14. 200406       5761 G 7614587.96 15402624.3  
  15. 200405       5763 J 7788036.32 15778496.6  
  16. 200406       5763 J 7990460.25 16852202.8  
  17. BILL_M  AREA_CODE N LOCAL_FARE   SUM_FARE  
  18. ------ ---------- - ---------- ----------  
  19. 200405       5763 G 8861742.59 17991798.1  
  20. 200406       5763 G 9130055.46 9130055.46  
  21. 13 rows selected.  

 

4.RANGE

RANGE是逻辑行，是按单元格值和偏移量计算窗口范围.

Range是逻辑行的范围 ，要经过 计算 的，一般range后面是数值或时间间隔等，这样根据 当行和range的表达 式能计算当

行对应的窗口范围；

[c-sharp:collapse] view plaincopyprint?

  1. select bill_month,area_code,net_type,local_fare,sum(local_fare) over(order by local_fare  range between current row and 122350 following) sum_fare  
  2.   from bill  
  3. /  
  4. BILL_M  AREA_CODE N LOCAL_FARE   SUM_FARE  
  5. ------ ---------- - ---------- ----------  
  6. 200405       5761 J 5667089.85 11371432.9  
  7. 200406       5761 J 5704343.05 5704343.05  
  8. 200405       5764 G 6028670.45 6028670.45  
  9. 200406       5762 J 6238068.05 6238068.05  
  10. 200406       5764 G 6387706.01 12846827.5  
  11. 200405       5764 J 6459121.49 13016114.1  
  12. 200406       5762 G  6556992.6  6556992.6  
  13. 200406       5764 J 6907481.66 6907481.66  
  14. 200406       5761 G 7614587.96 7614587.96  
  15. 200405       5763 J 7788036.32 7788036.32  
  16. 200406       5763 J 7990460.25 7990460.25  
  17. BILL_M  AREA_CODE N LOCAL_FARE   SUM_FARE  
  18. ------ ---------- - ---------- ----------  
  19. 200405       5763 G 8861742.59 8861742.59  
  20. 200406       5763 G 9130055.46 9130055.46  
  21. 13 rows selected.  

 

5.RATIO_TO_REPORT

The RATIO_TO_REPORT function computes the ratio of a value to the sum of a set of values. If the expression value expression evaluates to NULL , RATIO_TO_REPORT also evaluates to NULL , but it is treated as zero for computing the sum of values for the denominator. Its syntax is: 

RATIO_TO_REPORT ( expr ) OVER ( [query_partition_clause] )

[c-sharp:collapse] view plaincopyprint?

  1. select bill_month,area_code,net_type,local_fare,ratio_to_report(local_fare) over() rate  
  2. from bill  
  3. /  
  4. BILL_M  AREA_CODE N LOCAL_FARE       RATE  
  5. ------ ---------- - ---------- ----------  
  6. 200405       5761 J 5667089.85 .062047734  
  7. 200405       5763 G 8861742.59 .097025293  
  8. 200405       5763 J 7788036.32 .085269516  
  9. 200405       5764 G 6028670.45 .066006602  
  10. 200405       5764 J 6459121.49 .070719517  
  11. 200406       5761 G 7614587.96 .083370468  
  12. 200406       5761 J 5704343.05 .062455612  
  13. 200406       5762 G  6556992.6 .071791086  
  14. 200406       5762 J 6238068.05 .068299251  
  15. 200406       5763 G 9130055.46 .099962992  
  16. 200406       5763 J 7990460.25 .087485812  
  17. BILL_M  AREA_CODE N LOCAL_FARE       RATE  
  18. ------ ---------- - ---------- ----------  
  19. 200406       5764 G 6387706.01 .069937604  
  20. 200406       5764 J 6907481.66 .075628515  
Kenyataan：
Kandungan artikel ini disumbangkan secara sukarela oleh netizen, dan hak cipta adalah milik pengarang asal. Laman web ini tidak memikul tanggungjawab undang-undang yang sepadan. Jika anda menemui sebarang kandungan yang disyaki plagiarisme atau pelanggaran, sila hubungi admin@php.cn
Artikel sebelumnya：利用sqoop 将 hive/hdfs数据 导入 Oracle中Artikel seterusnya：Access入门教程 9.6 查询的数据表视图

Artikel berkaitan

Lihat lagi