SQL/DB2 替代扩展 CASE 语句

Question

I have a project where the data in single/multiple columns needs to be classified by group. My program uses a CASE statement to create more general categories. Since my code is quickly becoming unwieldy I don't think I have a very good approach.

How can I do better?

My Db2 query uses a 'CASE STATEMENT to generate the 'classification' column:

SELECT id
,plant
,CASE
     
when plant like 'rose'
     
 then 'flower'

when plant like 'wildflower'
 then 'flower'
when plant like 'spruce'
 then 'tree'
when plant like 'willow'
 then 'tree'
when plant like 'peony'
 then 'flower'
when plant like 'oak'
 then 'tree'
else 'unknown'
     
end as classification
FROM green_table

+----+------------+----------------+
| id | plant      | classification |
+----+------------+----------------+
| 1  | rose       | flower         |
+----+------------+----------------+
| 2  | wildflower | flower         |
+----+------------+----------------+
| 3  | spruce     | tree           |
+----+------------+----------------+
| 4  | willow     | tree           |
+----+------------+----------------+
| 5  | peony      | flower         |
+----+------------+----------------+
| 6  | oak        | tree           |
+----+------------+----------------+

Answer 1

Use a join to a derived table:

select g.*, coalesce(c.classification, 'unknown') as classification
from green_table g left join
     (select 'rose' as plant, 'flower' as classification from sysibm.sysdummy1 union all
      select 'wildflower' as plant, 'flower' as classification from sysibm.sysdummy1 union all
      . . .
     ) c
     on g.plant = c.plant;

I would actually recommend that you have a table called plants with one row per plant and the classification -- and a numeric primary key that you use to join to the name as well.

Note: This uses string equality rather than like . Equality seems sufficient and like will have a big impact on the performance of join .

Answer 2

You can use in operator for listing the same category

SELECT id,plant,
CASE   
  when plant in ('rose','wildflower','peony') then 'flower'
  when plant in ('spruce','willow','oak') then 'tree'
else 'unknown' end as classification
FROM green_table

Answer 3

You can use DECODE() . It even includes an "else" case. For example:

SELECT 
  id,
  plant,
  decode(plant, 'rose', 'flower', 'wildflower', 'flower', 'spruce', 'tree',
                'willow', 'tree', 'peony', 'flower', 'oak', 'tree',
                'unknown') as classification
FROM green_table