Oracle SQL将行转换为列

Question

我有以下数据结果集。

REQUEST_ID ATTRIBUTE_TYPE VENDOR LANG_ID PROJ_DESC  COST_TYPE VALUE 
     1 DOCUMENT  JLK 1188  SAMPLE PROJECT1 USAGE_COST 500 
     1 DOCUMENT  JLK 1188  SAMPLE PROJECT1 EXP_COST 350 
     1 DOCUMENT  JLK 1188  SAMPLE PROJECT1 COMMENTS OK 
     1 DOCUMENT  JLK 1194  SAMPLE PROJECT1 USAGE_COST 400 
     1 DOCUMENT  JLK 1194  SAMPLE PROJECT1 EXP_COST 575 
     1 DOCUMENT  JLK 1194  SAMPLE PROJECT1 COMMENTS DONE 
     1 DOCUMENT  JLK 1188  SAMPLE PROJECT2 USAGE_COST 100 
     1 DOCUMENT  JLK 1188  SAMPLE PROJECT2 EXP_COST 200 
     1 DOCUMENT  JLK 1188  SAMPLE PROJECT2 COMMENTS ACCEPTED 
     1 DOCUMENT  JLK 1194  SAMPLE PROJECT2 USAGE_COST 300 
     1 DOCUMENT  JLK 1194  SAMPLE PROJECT2 EXP_COST 400 
     1 DOCUMENT  JLK 1194  SAMPLE PROJECT2 COMMENTS GOOD 

我想调换使用列LANG_ID，PROJ_DESC，COST_TYPE和VALUE这些行。

我想达到以下结果：

REQUEST_ID ATTRIBUTE_TYPE VENDOR LANG_ID PROJ_DESC  USAGE_COST EXP_COST TOTAL_COST COMMENTS 
     1 DOCUMENT  JLK 1188 SAMPLE PROJECT1  500  350  850 OK 
     1 DOCUMENT  JLK 1194 SAMPLE PROJECT1  400  575  975 DONE 
     1 DOCUMENT  JLK 1188 SAMPLE PROJECT2  100  200  300 ACCEPTED 
     1 DOCUMENT  JLK 1194 SAMPLE PROJECT2  300  400  700 GOOD 

任何帮助就这将是高度赞赏。

谢谢。

Answer 1

一种方法是使用有条件聚集：

select REQUEST_ID, ATTRIBUTE_TYPE, VENDOR, LANG_ID, PROJ_DESC, 
     max(case when COST_TYPE = 'USAGE_COST' then value end) as usage_cost, 
     max(case when EXP_TYPE = 'EXP_COST' then value end) as exp_cost, 
     max(case when COST_TYPE = 'COMMENTS' then value end) as comments 
from t 
group by REQUEST_ID, ATTRIBUTE_TYPE, VENDOR, LANG_ID, PROJ_DESC; 

注意：您在VALUE列存储数字和字符串。当你想把这些值作为它们的本地类型时，这可能会很棘手。

Answer 2

使用PIVOT的解决方案（自Oracle 11.1起可用）。这增加了total_cost列，并明确地将成本转换为数字。

with 
    input_data(REQUEST_ID, ATTRIBUTE_TYPE, VENDOR, LANG_ID, PROJ_DESC, COST_TYPE, VALUE) as (
    select 1, 'DOCUMENT', 'JLK', 1188, 'SAMPLE PROJECT1', 'USAGE_COST', '500'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1188, 'SAMPLE PROJECT1', 'EXP_COST' , '350'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1188, 'SAMPLE PROJECT1', 'COMMENTS' , 'OK'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1194, 'SAMPLE PROJECT1', 'USAGE_COST', '400'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1194, 'SAMPLE PROJECT1', 'EXP_COST' , '575'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1194, 'SAMPLE PROJECT1', 'COMMENTS' , 'DONE'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1188, 'SAMPLE PROJECT2', 'USAGE_COST', '100'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1188, 'SAMPLE PROJECT2', 'EXP_COST' , '200'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1188, 'SAMPLE PROJECT2', 'COMMENTS' , 'ACCEPTED' from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1194, 'SAMPLE PROJECT2', 'USAGE_COST', '300'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1194, 'SAMPLE PROJECT2', 'EXP_COST' , '400'  from dual union all 
    select 1, 'DOCUMENT', 'JLK', 1194, 'SAMPLE PROJECT2', 'COMMENTS' , 'GOOD'  from dual 
    ) 
select request_id, attribute_type, vendor, lang_id, proj_desc, 
     to_number(usage_cost) as usage_cost, to_number(exp_cost) as exp_cost, 
     to_number(usage_cost) + to_number(exp_cost) as total_cost, comments 
from input_data 
pivot (max(value) for cost_type in ('USAGE_COST' as usage_cost, 
             'EXP_COST' as exp_cost , 
             'COMMENTS' as comments 
            ) 
    ) 
; 

输出：

REQUEST_ID ATTRIBUT VEN LANG_ID PROJ_DESC  USAGE_COST EXP_COST TOTAL_COST COMMENTS 
---------- -------- --- ------- --------------- ---------- ---------- ---------- -------- 
     1 DOCUMENT JLK 1194 SAMPLE PROJECT2  300  400  700 GOOD 
     1 DOCUMENT JLK 1194 SAMPLE PROJECT1  400  575  975 DONE 
     1 DOCUMENT JLK 1188 SAMPLE PROJECT1  500  350  850 OK 
     1 DOCUMENT JLK 1188 SAMPLE PROJECT2  100  200  300 ACCEPTED