查询以基于其他列获取最接近的匹配值

Question

Based on Table 1 below, I am trying to write a query that allows me to create a new table, fill up the empty companyID with closest match based on City, Country and Region.

Table 1 is the sample table, whereas Table 2 is the expected output.

Table 1

Cust | City | Country | Region | CompanyID
ABC  |   KL |      MY |   APAC | 123456
ABC  |   KL |      MY |   APAC | 
ABC  |   JB |      MY |   APAC | 
ABC  |   SY |      AU |   APAC | 778899
ABC  |   ME |      AU |   APAC | 
GHI  |   DB |      AE |   EMEA | 112233
GHI  |   AD |      AE |   EMEA |

Table 2

Cust | City | Country | Region | CompanyID
ABC  |   KL |      MY |   APAC | 123456
ABC  |   KL |      MY |   APAC | 123456
ABC  |   JB |      MY |   APAC | 123456
ABC  |   SY |      AU |   APAC | 778899
ABC  |   ME |      AU |   APAC | 778899
GHI  |   DB |      AE |   EMEA | 112233
GHI  |   AD |      AE |   EMEA | 112233

Feels a bit like a VLOOKUP thing but not sure how to start. Will sincerely appreciate any help please.

Thanks!

Answer 1

In a subquery.or using with create inline views of all defaults. Then outer join to those defaults.

I first created a table called usr_test_table which has same data as mentioned above. And note that my database is oracle so do not get confused by "from dual". That is a special table in oracle with one row in one column.

select t1.cust, t1.city, t1.country, t1.region,
       case 
           when t1.companyid is not null then t1.companyid
           else (
                   select max(companyid)
                   from usr_test_table t2
                   where t2.country = t1.country
               )
       end companyid
from usr_test_table t1

It is slightly simpler than earlier comment about outer join to an inline view. Since now databases allow a subquery in select clause such situatioins are simpler to handle that way.