动态处理选择条件

Question

我有一个类似的程序

Procedure P_GENDER_REP
 (
  l_province  IN  varchar,
  l_district  IN  varchar,
  l_village   IN  varchar,
  l_tribe     IN  varchar,
  l_clan      IN  varchar,
  l_refcursor out sys_refcursor
 )

我在这样的过程中有一个选择查询

select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
from itaukei_data_store b,itaukei_data_store_key a 
where a.reference_no = b.reference_no and a.record_no = b.record_no and a.province = l_province 
  and a.district = l_district and a.village = l_village and a.tribe= l_tribe 
    and a.clan = l_clan      
order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;

现在我需要一个选择查询来处理给定的输入。例如，如果只给出 l_province，我需要动态忽略所有其他 where 条件。请帮帮我。

Answer 1

通常您通过在每个条件中添加OR l_variable IS NULL 来做到这一点：

select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
from itaukei_data_store b INNER JOIN itaukei_data_store_key a 
ON a.reference_no = b.reference_no 
  and a.record_no = b.record_no 
  and a.province = l_province 
WHERE l_district IS NULL OR a.district = l_district
  and l_village IS NULL OR a.village = l_village
  and l_tribe IS NULL OR a.tribe= l_tribe
  and l_clan IS NULL OR a.clan = l_clan
order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;

但是，这有时可能会影响性能。对每种组合使用 IF 会更好。或者，您可以尝试使用支持短路的 CASE 语句，如下所示：

select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
from itaukei_data_store b INNER JOIN itaukei_data_store_key a 
ON a.reference_no = b.reference_no 
  and a.record_no = b.record_no 
  and a.province = l_province 
WHERE a.district = CASE WHEN l_district IS NULL THEN a.district ELSE l_district END
  and a.village = CASE WHEN l_village IS NULL THEN a.village ELSE l_village END
  and a.tribe = CASE WHEN l_tribe IS NULL THEN a.tribe ELSE l_tribe END
  and a.clan = CASE WHEN l_clan IS NULL THEN a.clan ELSE l_clan END
order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;

你有第三种选择。您可以使用COALESCE 代替CASE，如下所示：

select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
from itaukei_data_store b INNER JOIN itaukei_data_store_key a 
ON a.reference_no = b.reference_no 
  and a.record_no = b.record_no 
  and a.province = l_province 
WHERE a.district = COALESCE(l_district, a.district)
  and a.village = COALESCE(l_village, a.village)
  and a.tribe = COALESCE(l_tribe, a.tribe)
  and a.clan = COALESCE(l_clan, a.clan)
order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;

尝试每一种，看看哪一种表现更好。顺便说一句，我还将您的 FROM 子句从逗号分隔的表列表更改为标准的 INNER JOIN。请养成使用 JOIN 的习惯，因为它现在被认为是“最佳选择”

Answer 2

给出的示例是 SQL Server， 第一种方法，简单的如果

Declare @l_province int
Declare @l_district int
Declare @l_village int

If @l_province is not null and @l_district is null  and @l_village is null -- Only l_province is given
Begin
    select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
        from itaukei_data_store b,itaukei_data_store_key a 
    order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;
End
Else
Begin
    select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
        from itaukei_data_store b,itaukei_data_store_key a 
    where a.reference_no = b.reference_no 
            and a.record_no = b.record_no 
            and a.province = l_province 
            and a.district = l_district 
            and a.village = l_village 
            and a.tribe= l_tribe 
            and a.clan = l_clan      
    order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;
End

---------------------使用 OR , AND 的第二种方式

select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data 
        from itaukei_data_store b,itaukei_data_store_key a 
    where (
            (
                (@l_province is not null and @l_district is null  and @l_village is null) -- Only l_province is given
                AND
                (1=1)
            )
            OR
            (
                (@l_province is not null and (@l_district is not null  OR @l_village is not null))
                AND
                (
                        a.reference_no = b.reference_no 
                        and a.record_no = b.record_no 
                        and a.province = l_province 
                        and a.district = l_district 
                        and a.village = l_village 
                        and a.tribe= l_tribe 
                        and a.clan = l_clan 
                )
            )
            OR
            (
                (@l_province is null)
                AND
                (
                        a.reference_no = b.reference_no 
                        and a.record_no = b.record_no 
                        and a.province = l_province 
                        and a.district = l_district 
                        and a.village = l_village 
                        and a.tribe= l_tribe 
                        and a.clan = l_clan 
                )

            )
        )   
    order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;

其实为了简化存储过程，我建议你带一个参数来判断where子句是否应用！为了计算该变量，您需要在代码后面或存储过程开始时编写一些代码。它会让生活变得轻松。

Answer 3

只是为了添加另一个技巧。

如果每个列的值不同，您可以使用coalesce

select  a.province,a.district,a.village,a.tribe,a.clan,b.cdr_data 
from    itaukei_data_store b,itaukei_data_store_key a 
where   a.reference_no = b.reference_no 
and     a.record_no = b.record_no 
and     coalesce(a.province,a.district,a.village,a.tribe,a.clan) = 
        coalesce(l_province,l_district,l_village,l_tribe,l_clan)
order by a.district,a.village,a.tribe,a.clan;

合并返回列表中的第一个非空值

Answer 4

您可以对参数进行 nvl，以便在设置它们时使用它们，否则将采用比较表中的匹配值..应该具有预期的效果：

select a.province,a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN,b.cdr_data
from itaukei_data_store b,itaukei_data_store_key a 
where a.reference_no = b.reference_no 
and a.record_no = b.record_no 
and a.province = nvl(l_province, a.province)
and a.district = nvl(l_district, a.district)
and a.village = nvl(l_village, a.village)
and a.tribe = nvl(l_tribe, a.tribe) 
and a.clan = nvl(l_clan, a.clan)
order by a.DISTRICT,a.VILLAGE,a.TRIBE,a.CLAN;