Table function

Question

Hi,
1 . what is table()?
2. Can you please explain the below line of code?
PROCEDURE ins_alias (
pv_app_o OUT t$app.app%TYPE,
pv_name_i IN t$app.name%TYPE,
pva_apps_i IN id_vt
AS
lv_app t$app.app%TYPE;
lv_userid t$app.userid%TYPE;
lva_new_apps t$app.apps%TYPE;
BEGIN
-- Only add valid apps (ones that exist and are not user-created)
SELECT t1.COLUMN_VALUE BULK COLLECT INTO lva_new_apps
FROM TABLE (pva_apps_i) t1
WHERE EXISTS
(SELECT 1
FROM t$app a1
WHERE a1.app = t1.column_value AND
a1.userid IS NULL)
ORDER BY COLUMN_VALUE;
what is column_value and table() in the code?
Thanks,

Answer 1

table() takes a certain sort of parameter and turns it into a table that can be queried.
Such parameters may be Pipelined Functions, or certain types of collections/arrays.
So, in your example:
pva_apps_i is of a type id_vt, which I'm guessing is some sort of collection type (sql table type).
COLUMN_VALUE is the default name given to a TABLE() function's returned column, if there is only one column returned, indicating your collection is a simple array with just a list of values.
It's possible to create a collection of object types and have the table function return the column names as the names of the attributes of the object, such as the following example which uses a pipelined function...
CREATE OR REPLACE TYPE myrec AS OBJECT
( col1   VARCHAR2(10),
  col2   VARCHAR2(10)
CREATE OR REPLACE TYPE myrectable AS TABLE OF myrec
CREATE OR REPLACE FUNCTION pipedata(p_str IN VARCHAR2) RETURN myrectable PIPELINED IS
  v_str VARCHAR2(4000) := REPLACE(REPLACE(p_str, '('),')');
  v_obj myrec := myrec(NULL,NULL);
BEGIN
  LOOP
    EXIT WHEN v_str IS NULL;
    v_obj.col1 := SUBSTR(v_str,1,INSTR(v_str,',')-1);
    v_str := SUBSTR(v_str,INSTR(v_str,',')+1);
    IF INSTR(v_str,',')>0 THEN
      v_obj.col2 := SUBSTR(v_str,1,INSTR(v_str,',')-1);
      v_str := SUBSTR(v_str,INSTR(v_str,',')+1);
    ELSE
      v_obj.col2 := v_str;
      v_str := NULL;
    END IF;
    PIPE ROW (v_obj);
  END LOOP;
  RETURN;
END;
SQL> select *
  2  from table(pipedata('(1,2),(3,4),(5,6)'));
COL1       COL2
1          2
3          4
5          6