Question

I am doing a query that finds all cats born between two dates. Each cat has a name, or multiple names.

The initial query is something like

SELECT Id, Color FROM Cat WHERE Cat.BirthDate > dat_min AND Cat.BirthDate < dat_max;

I also have a table called CatName which for each Cat Id, has one or more names that this cat has been given by its different owners. I only want to return the first name that matches the Id in the CatName table, as part of the query. So something like:

SELECT Id, Color, Name FROM Cat JOIN CatName on .....

for a cat that has 5 names, will return 5 rows. I only want one row, the first one. If I was only retrieving data for one cat, then I would just use ROWNUM to limit it to 1 query, but I am trying to get a list of all cats including their name, so I can’t do this.

Can anyone offer some guidance? I guess it doesn’t have to be plsql specific, the technique will be the same I imagine.

Answer 1

There are several methods.

1. You could use an inline query:

   SELECT id, color, 
          (SELECT name FROM CatName cn WHERE cn.id = c.id AND ROWNUM = 1) Name
     FROM cat c
    WHERE ...
   

2. You could use a join then analytics:

   SELECT id, color, Name
     FROM (SELECT Id, Color, Name,
                  row_number() OVER (PARTITION BY id ORDER BY 1) rn
             FROM Cat JOIN CatName on .....)
    WHERE rn = 1
   

3. You can use an aggregate:

   SELECT id, color, MAX(name) name
     FROM Cat JOIN CatName on .....
    GROUP BY id, color
   

From a performance point of view, assuming that CatName is indexed by CatId:

• if the number of cats returned is smallish, or you only want the very first few cats among many, solution 1 can be really fast,
• if the dataset returned is large and you want all cats, then solution 2 and 3 can make good use of the efficient HASH JOIN.