Correctly use collection method

Question

CREATE  OR  REPLACE  TYPE  nvarchar2_list_type AS TABLE OF NVARCHAR2(100);


CREATE   TABLE test_table(
   id number primary key,
   cars_list     nvarchar2_list_type 
)
NESTED TABLE cars_list  STORE AS cars_list_storage_table;


insert into test_table(id, cars_list)
    values(1,  nvarchar2_list_type( 'AUDI', 'MERCEDES')  );

All above operations completed success, 1 rows inserted in table test_table, now i write this function:

create or replace function get_cnt 
return number
as
  ret_val number;
begin
  SELECT  cars_list.COUNT    
    INTO  ret_val
   from test_table where id  = 1;

   return ret_val;
end;

This gives error: ORA-00904: "CARS_LIST"."COUNT": invalid identifier

Tell please what is wrong here?

As I know, COUNT method must be used just so (from here)

Answer 1

No, you cannot use count method in this situation. You have SQL nested table at hand, count method is used only with PL/SQL collections.

To count number of nested table's elements you can either unnest that nested table or use scalar sub-query:

Unnesting:

SQL> select id
  2       , count(*) as cnt
  3     from test_table t
  4     cross join table(t.cars_list)
  5    group by id
  6  ;

        ID        CNT
---------- ----------
         1          2

Scalar sub-query:

SQL> select id
  2       , (select count(column_value)
  3            from table(t.cars_list)) as cnt
  4     from test_table t
  5  ;
        ID        CNT
---------- ----------
         1          2

Answer 2

Use

  Select
  Cardinality(cars_list) from test_table

Answer 3

I can't explain why that doesn't work, but this does:

select (select count(*) from table(cars_list))
into ret_val
from test_table
where id  = 1;

Answer 4

Oracle is expecting column name or function in its select list, but what you are giving is collection build in method that operates on collections only.

You can achieve the same using scalar sub query

SELECT (select count(1) from table(cars_list)) as "COUNT" 
FROM test_table 
WHERE id = 1;