如何在沒有聚合的情況下將這些行轉換爲Oracle中的列？

Question

大家好我有下表，它是由一個可怕的開發人員設計的

該表包含USERS指紋和他們的USERID的數據。 的FINGERID值從（1,2,3,4,5,6,7,8,9,10,11,12,13,14）的ISLOCKEDISDUPLICATE和值是1或0 ...每個FINGERID已存儲在BLOB

Name  Null  Type   
----------- -------- ------------- 
ROW_ID    NUMBER   
USERID  NOT NULL NVARCHAR2(75) 
TEMPLATE    BLOB   
FINGERID NOT NULL NVARCHAR2(20) 
ISLOCKED    NUMBER(3)  
ISDUPLICATE   NUMBER(3) 

有像我想要顯示它，因爲2周水平的方式，但沒有運氣...以下是我的select語句

Name     Null  Type   
----------------------- -------- -------------   
USERID     NOT NULL NVARCHAR2(75) 
FINGERID_1      NVARCHAR2(20) 
FINGERID_1_TEMPLATE    BLOB   
FINGERID_1_ISLOCKED    NUMBER(3)  
FINGERID_1_ISDUPLICATE   NUMBER(3)  
FINGERID_2      NVARCHAR2(20) 
FINGERID_2_TEMPLATE    BLOB   
FINGERID_2_ISLOCKED    NUMBER(3)  
FINGERID_2_ISDUPLICATE   NUMBER(3)  
FINGERID_3      NVARCHAR2(20) 
FINGERID_3_TEMPLATE    BLOB   
FINGERID_3_ISLOCKED    NUMBER(3)  
FINGERID_3_ISDUPLICATE   NUMBER(3)  
FINGERID_4      NVARCHAR2(20) 
FINGERID_4_TEMPLATE    BLOB   
FINGERID_4_ISLOCKED    NUMBER(3)  
FINGERID_4_ISDUPLICATE   NUMBER(3)  
FINGERID_5      NVARCHAR2(20) 
FINGERID_5_TEMPLATE    BLOB   
FINGERID_5_ISLOCKED    NUMBER(3)  
FINGERID_5_ISDUPLICATE   NUMBER(3)  
FINGERID_6      NVARCHAR2(20) 
FINGERID_6_TEMPLATE    BLOB   
FINGERID_6_ISLOCKED    NUMBER(3)  
FINGERID_6_ISDUPLICATE   NUMBER(3)  
FINGERID_7      NVARCHAR2(20) 
FINGERID_7_TEMPLATE    BLOB   
FINGERID_7_ISLOCKED    NUMBER(3)  
FINGERID_7_ISDUPLICATE   NUMBER(3)  
FINGERID_8      NVARCHAR2(20) 
FINGERID_8_TEMPLATE    BLOB   
FINGERID_8_ISLOCKED    NUMBER(3)  
FINGERID_8_ISDUPLICATE   NUMBER(3)  
FINGERID_9      NVARCHAR2(20) 
FINGERID_9_TEMPLATE    BLOB   
FINGERID_9_ISLOCKED    NUMBER(3)  
FINGERID_9_ISDUPLICATE   NUMBER(3)  
FINGERID_10      NVARCHAR2(20) 
FINGERID_10_TEMPLATE    BLOB   
FINGERID_10_ISLOCKED    NUMBER(3)  
FINGERID_10_ISDUPLICATE   NUMBER(3) 
FINGERID_11      NVARCHAR2(20) 
FINGERID_11_TEMPLATE    BLOB   
FINGERID_11_ISLOCKED    NUMBER(3)  
FINGERID_11_ISDUPLICATE   NUMBER(3) 

我曾嘗試使用這種期望的結構代碼但沒有運氣，因爲我無法顯示TEMPLATE（BLOB），ISLOCKED，ISDUPLICATED 每個FINGERID，因爲它不會與group by工作，pivot是不是一個很好的選擇，因爲它需要聚集......我只是想要顯示的數據只是它是如何的，但在每個水平方式用戶名 。

`select 
     USERID, 
     SUM(case when FINGERID=1 then 1 END) as FINGERID_1, 
     SUM(case when FINGERID=2 then 2 END) as FINGERID_2, 
     SUM(case when FINGERID=3 then 3 END) as FINGERID_3, 
     SUM(case when FINGERID=4 then 4 END) as FINGERID_4, 
     SUM(case when FINGERID=5 then 5 END) as FINGERID_5, 
     SUM(case when FINGERID=6 then 6 END) as FINGERID_6, 
     SUM(case when FINGERID=7 then 7 END) as FINGERID_7, 
     SUM(case when FINGERID=8 then 8 END) as FINGERID_8, 
     SUM(case when FINGERID=9 then 9 END) as FINGERID_9, 
     SUM(case when FINGERID=10 then 10 END) as FINGERID_10, 
     SUM(case when FINGERID=11 then 11 END) as FINGERID_11, 
     SUM(case when FINGERID=12 then 12 END) as FINGERID_12, 
     SUM(case when FINGERID=13 then 13 END) as FINGERID_13, 
     SUM(case when FINGERID=14 then 14 END) as FINGERID_14 
     FROM TBLUSERFINGER group by userid;` 

Answer 1

這些數據類型確實存在問題。相反，您可以多次使用LEFT JOIN：

select u.userid, 
     f1.fingerid as fingerid_1, 
     f1.template as fingerid_1_template, 
     f1.islocked as fingerid_1_islocked, 
     f1.isduplicate as fingerid_1_isduplicate, 
     . . . 
from (select distinct userid from TBLUSERFINGER) u left join 
    TBLUSERFINGER f1 
    on f1.userid = u.userid and f1.fingerid = 1 left join 
    TBLUSERFINGER f2 
    on f2.userid = u.userid and f2.fingerid = 2 left join 
    . . . 
    TBLUSERFINGER f14 
    on f14.userid = u.userid and f14.fingerid = 14 ; 

Answer 2

下面是另一種方法。如果你的桌子非常大，並且多次連接到它自己的速度太慢，你可以將除blob之外的所有東西都擺出來，將rowid存儲在可以找到blob的地方。例如：

create table testx (
ROW_ID    NUMBER,   
USERID  NVARCHAR2(75) NOT NULL, 
TEMPLATE    BLOB  ,  
FINGERID  NVARCHAR2(20) NOT NULL, 
ISLOCKED    NUMBER(3)  , 
ISDUPLICATE   NUMBER(3) 
); 

create unique index testx_uidx on testx(userid, fingerid); 

-- insert some test data 
insert into testx values (1,'UID1',RAWTOHEX(UTL_RAW.CAST_TO_RAW('LOB1')),'1',1,0); 
insert into testx values (2,'UID1',RAWTOHEX(UTL_RAW.CAST_TO_RAW('LOB2')),'2',1,0); 
insert into testx values (3,'UID1',RAWTOHEX(UTL_RAW.CAST_TO_RAW('LOB3')),'3',0,0); 
insert into testx values (4,'UID2',RAWTOHEX(UTL_RAW.CAST_TO_RAW('LOB1')),'1',1,0); 
insert into testx values (5,'UID2',RAWTOHEX(UTL_RAW.CAST_TO_RAW('LOB2')),'2',1,0); 
insert into testx values (6,'UID2',RAWTOHEX(UTL_RAW.CAST_TO_RAW('LOB3')),'3',0,0); 
commit; 

-- pivot out what we can, show rowid to blob data 
select * from (
    select userid, rowid as template_rowid, fingerid,islocked,isduplicate 
    from testx 
) 
pivot ( 
    max(template_rowid) rowid_of_template, max(islocked) mislocked, max(isduplicate) misduplicate 
    for fingerid in ('1' fid1,'2' fid2,'3' fid3,'4' fid4,'5' fid5,'6' fid6,'7' fid7,'8' fid8,'9' fid9,'10' fid10,'11' fid11,'12' fid12,'13' fid13,'14' fid14) 
); 

輸出：

USERID,FID1_ROWID_OF_TEMPLATE,FID1_MISLOCKED,FID1_MISDUPLICATE,FID2_ROWID_OF_TEMPLATE,FID2_MISLOCKED,FID2_MISDUPLICATE,FID3_ROWID_OF_TEMPLATE,FID3_MISLOCKED,FID3_MISDUPLICATE,FID4_ROWID_OF_TEMPLATE,FID4_MISLOCKED,FID4_MISDUPLICATE,FID5_ROWID_OF_TEMPLATE,FID5_MISLOCKED,FID5_MISDUPLICATE,FID6_ROWID_OF_TEMPLATE,FID6_MISLOCKED,FID6_MISDUPLICATE,FID7_ROWID_OF_TEMPLATE,FID7_MISLOCKED,FID7_MISDUPLICATE,FID8_ROWID_OF_TEMPLATE,FID8_MISLOCKED,FID8_MISDUPLICATE,FID9_ROWID_OF_TEMPLATE,FID9_MISLOCKED,FID9_MISDUPLICATE,FID10_ROWID_OF_TEMPLATE,FID10_MISLOCKED,FID10_MISDUPLICATE,FID11_ROWID_OF_TEMPLATE,FID11_MISLOCKED,FID11_MISDUPLICATE,FID12_ROWID_OF_TEMPLATE,FID12_MISLOCKED,FID12_MISDUPLICATE,FID13_ROWID_OF_TEMPLATE,FID13_MISLOCKED,FID13_MISDUPLICATE,FID14_ROWID_OF_TEMPLATE,FID14_MISLOCKED,FID14_MISDUPLICATE 
"UID1","AAzWrKACUAAA1ffAAA",1,0,"AAzWrKACUAAA1ffAAB",1,0,"AAzWrKACUAAA1ffAAC",0,0,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 
"UID2","AAzWrKACUAAA1ffAAD",1,0,"AAzWrKACUAAA1ffAAE",1,0,"AAzWrKACUAAA1ffAAF",0,0,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 

應用程序然後可以非常迅速地進行查找（以顯示例如團塊圖像）根據需要使用的rowid。