oracle sql对用例条件的澄清在where子句中，如果没有索引，则无法计算错误sdo_nn_Sql_Oracle

oracle sql对用例条件的澄清在where子句中，如果没有索引，则无法计算错误sdo_nn

sql oracle

oracle sql对用例条件的澄清在where子句中，如果没有索引，则无法计算错误sdo_nn,sql,oracle,Sql,Oracle,我正在使用oracle 11g和oracle spatial，如果我在where条件下使用以下距离，则此查询有效 mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='|| c1.tier_radius||' unit=KM', 1) = 'TRUE' 但是，我希望根据条件计算距离。下面的查询给出了错误sdo_nn在没有索引的情况下无法计算，我希望根据类似这样的条件计算距离 select

我正在使用oracle 11g和oracle spatial，如果我在where条件下使用以下距离，则此查询有效

mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='|| 
c1.tier_radius||' unit=KM', 1) = 'TRUE'

但是，我希望根据条件计算距离。下面的查询给出了错误sdo_nn在没有索引的情况下无法计算，我希望根据类似这样的条件计算距离

select
        /*+ first_rows LEADING(c) USE_NL(c1 b1) INDEX(b1 BRANCH_LOCATION_SPIX)*/
        b1.BRANCH_ID,
        mdsys.sdo_nn_distance(1) branchDistance
      from
        branch b1, city c1
      where
        mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='|| (case  when b1.restaurant_id not in (3970,3971,3972) then c1.tier_radius
        else 15 end) ||' unit=KM', 1) = 'TRUE'   
        and b1.city_id = c1.city_id

我们知道那部分

mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='|| 
c1.tier_radius||' unit=KM', 1) = 'TRUE'

如果有效，那么我认为您可以使用这样一个包含

union all

的查询，如下所示：

select  /*+ first_rows LEADING(c) USE_NL(c1 b1) INDEX(b1 BRANCH_LOCATION_SPIX)*/
        b1.BRANCH_ID,
        mdsys.sdo_nn_distance(1) branchDistance
  from branch b1 
  join city c1 on c1.city_id = b1.city_id
 where mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='||c1.tier_radius||' unit=KM', 1) = 'TRUE'
   and b1.restaurant_id not in (3970,3971,3972)
union all
select  /*+ first_rows LEADING(c) USE_NL(c1 b1) INDEX(b1 BRANCH_LOCATION_SPIX)*/
        b1.BRANCH_ID,
        mdsys.sdo_nn_distance(1) branchDistance
  from branch b1 
  join city c1 on c1.city_id = b1.city_id
 where mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance=15 unit=KM', 1) = 'TRUE'
   and b1.restaurant_id in (3970,3971,3972)

我们知道那部分

mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='|| 
c1.tier_radius||' unit=KM', 1) = 'TRUE'

如果有效，那么我认为您可以使用这样一个包含

union all

的查询，如下所示：

select  /*+ first_rows LEADING(c) USE_NL(c1 b1) INDEX(b1 BRANCH_LOCATION_SPIX)*/
        b1.BRANCH_ID,
        mdsys.sdo_nn_distance(1) branchDistance
  from branch b1 
  join city c1 on c1.city_id = b1.city_id
 where mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance='||c1.tier_radius||' unit=KM', 1) = 'TRUE'
   and b1.restaurant_id not in (3970,3971,3972)
union all
select  /*+ first_rows LEADING(c) USE_NL(c1 b1) INDEX(b1 BRANCH_LOCATION_SPIX)*/
        b1.BRANCH_ID,
        mdsys.sdo_nn_distance(1) branchDistance
  from branch b1 
  join city c1 on c1.city_id = b1.city_id
 where mdsys.sdo_nn ( b1.location, l_location , 'SDO_BATCH_SIZE=100 distance=15 unit=KM', 1) = 'TRUE'
   and b1.restaurant_id in (3970,3971,3972)

是的，union all工作正常，但不会出现性能问题，因为union正在执行两个查询而不是一个查询？@Ayub我从未遇到与

union all