How to return a resultset in a stored procedure without using Cursor

Question

Dear all,
I need to return a resultset in a stored procedure.
I know I can return a Cursor.
But because there are many complicate logics I need to carry out,
so I can finish all the logic in a SINGLE SQL select statement, so I think it can't
use Cursor as return value.
Does anybody know other approaches? Please help.
Thanks!

Answer 1

Some basic techniques that one can use in SQL. These are merely to illustrate (simplistically) different approaches.
> field3 = select count(*) from table2,table1 where table2.id = table1.id
This can be done via an analytical function (refer to the SQL Reference) or very simply, as an in-line select as follows:
SQL> select
2 o.object_name,
3 (select count(*) from user_objects u where u.object_type = o.object_type) as "ONE OF",
4 o.object_type
5 from user_objects o
6 where o.object_name = 'LISTFILES';
OBJECT_NAME ONE OF OBJECT_TYPE
LISTFILES 8 PROCEDURE
So LISTFILE is "one of 8" procedures. I would however rather so this via an analytic function as these are a lot more flexible. Note the WITH clause that allows "modularisation" of SQL - kind of like creating "sub-SQL results" in a single SQL statement:
SQL> with DATASET1 as(
2 select
3 object_name,
4 count(distinct object_name) over (partition by object_type) as TOTAL_TYPES,
5 object_type
6 from user_objects
7 ),
8 DATASET2 as(
9 select
10 name,
11 text as SOURCE_LINE,
12 line as LINE_NO,
13 count(line) over (partition by name) as TOTAL_LINES
14 from user_source
15 )
16 select
17 object_name,
18 '1 of '||total_types||' '||object_type as "TYPE",
19 source_line as "1st LINE",
20 total_lines as "TOTAL LINES"
21 from DATASET1 ds1,
22 DATASET2 ds2
23 where ds1.object_name = ds2.name
24 and ds1.object_name = 'LISTFILES'
25* and ds2.line_no = 1
SQL> /
OBJECT_NAME TYPE 1st LINE TOTAL LINES
LISTFILES 1 of 8 PROCEDURE procedure ListFiles( cDirectory in varchar2 ) 3
SQL>
SQL is quite powerful and analytical functions allows all kinds of aggregation processing per row, accessing the leading or lagging rows' data, etc.
You need to play around with this to get to grips with it and how to apply it.