如何存储一次交易中购买的多件商品？

Question

我是MySQL新手。我的数据库是客户购买的零售产品或webview数据。示例：

ID  Records
R1  {iphone5s,levis,Dell,Dior,amazon}
R2  {levis,Dell,amazon,Starbucks,google glass}
R3  {iphone5s,Dell,Dior,google glass}
R4  {iphone5s,levis,Starbucks}
R5  {iphone5s,levis,Dell,google glass}

我想将此数据存储到数据库中。我将此数据存储为：

ID    iphone5s   levis   Dell    Dior   amazon  Starbucks   google glass

R1       1         1      1        1       1         0          0

R2       0         1      1        0       1         1          1

R3       1         0      1        1       0         0          1

R4       1         1      0        0       0         1          0

R5       1         1      1        0       0         0          1


create table retail(ID varchar(50),
                    iphone5s int,
                    levis int,
                    Dell int,
                    Dior int,
                    amazon int,
                    Starbucks int,
                    googleglass int);

insert into retail
values ('r1',1,1,1,1,1,0,0), ('r2',0,1,1,0,1,1,1);

insert into retail
values ('r3',1,0,1,1,0,0,1),('r4',1,1,0,0,0,1,0),('r5',1,1,1,0,0,0,1);

现在我想检索具有iphone5s的ID并存储ID。类似，我想收集具有每个列名的 id 并存储相应的 id。

使用java我通过以下代码收集列名称：

ResultSet rs = stmt.executeQuery("SELECT * FROM retail");
ResultSetMetaData rsmd = rs.getMetaData();
int columnCount = rsmd.getColumnCount();
for (int i = 2; i < columnCount + 1; i++ ) {
  String name = rsmd.getColumnName(i);
}

现在我想从零售中选择ID

where columnname=1

。

列名每次都应该改变。

Answer 1

来自对问题的评论：

还有其他更好的方式将上述原始数据存储在数据库系统中吗？

当然。您应该将购买的每件商品存储在单独的行中，如下所示：

ID  item
--  ------------
R1  iphone5s
R1  levis
R1  Dell
R1  Dior
R1  amazon
R2  levis
R2  Dell
R2  amazon
R2  Starbucks
R2  google glass
...

这使得查询更容易（因为列名是不变的），并且如果您需要添加新产品，则无需更改数据库结构。您始终可以将数据“转换”为您现在拥有的格式，但您不应该以这种方式存储它。

Answer 2

您可以在 MySQL 中使用

UNPIVOT

等效项。现在您可以简单地使用

WHERE brand = 'name' AND is_available = 1

:

进行查询

SqlFiddle演示

create table retail(ID varchar(50),
                    iphone5s int,
                    levis int,
                    Dell int,
                    Dior int,
                    amazon int,
                    Starbucks int,
                    googleglass int);

insert into retail
values ('r1',1,1,1,1,1,0,0), ('r2',0,1,1,0,1,1,1),
       ('r3',1,0,1,1,0,0,1),('r4',1,1,0,0,0,1,0),('r5',1,1,1,0,0,0,1);

select id
from
(
  select
    t.id,
    c.brand,
    case c.brand
      when 'iphone5s'    then iphone5s
      when 'levis'       then levis
      when 'Dell'        then Dell
      when 'Dior'        then Dior
      when 'amazon'      then amazon
      when 'Starbucks'   then Starbucks
      when 'googleglass' then googleglass 
    end as is_available
  from retail t
  cross join
  (
    select 'iphone5s' as brand
    union all
    select 'levis'
    union all
    select 'Dell'
    union all
    select 'Dior'
    union all 
    select 'amazon'
    union all
    select 'Starbucks'
    union all
    select 'googleglass'
  ) c
) temp
where brand = 'iphone5s'   /* Here your brand/product_name */
  and is_available = 1
order by id;              /* If needed */

SqlFiddleDemo_UNPIVOT_TSQL

WITH cte AS
(
  SELECT ID, Brand, Val
  FROM
  (
     SELECT Id, iphone5s, levis, Dell, Dior, amazon, Starbucks, googleglass
     FROM retail
   ) t
   UNPIVOT
   (
     Val FOR Brand IN (iphone5s, levis, Dell,
                    Dior, amazon, Starbucks, googleglass)
    ) AS mrks
)
SELECT ID
FROM cte
WHERE Brand = 'Dior' 
AND [val] = 1;