Nested table variable intersect

Question

Hi everyone,
I've created several nested tables of records, and I'd like to find the intersection of these variables. Just to use a a simple example:
declare
  type test_record is record (
    col1 NUMBER,
    col2 VARCHAR2(255)
  type n_table is table of test_record;
  t1 n_table;
  t2 n_table;
  t3 n_table;
begin
  t1(1).col1 := 1;
  t1(1).col2 := 'A';
  t1.extend;
  t1(2).col1 := 2;
  t1(2).col2 := 'B';
  t2(1).col1 := 1;
  t2(1).col2 := 'A';
  t2.extend;
  t2(2).col1 := 2;
  t2(2).col2 := 'B';
end;Now what I'd like to do is find the intersection of tables t1 and t2, and store it in t3. I've read about using multiset operators, but those only seem to work on nested tables of scalar types, not records. Is there any way I could do this? If it helps, I am only considering one column in my intersection, so if I could somehow select col1 from t1 and col1 from t2, and find the intersection of that sub-set, that might work too.
I'd also like to do it without looping through every row in every table, since the number of tables and number of rows can be quite large.
Thanks!

Answer 1

Hey Gerard, thanks for the reply.
The collections only exist in memory, not in the database. I'm running the same raw data through a series of different functions, each of which returns a collection with some rows removed. My desired output is the set of rows which exist in every collection, so the intersection.
I'm hesitant to change the type from a record to an object since the type is declared in a package and there's a lot of other procedures that use it. But you did get me thinking about using straight SQL. So I've made a table to use for temporary storage,
CREATE TABLE TEMP_STORAGE  (
    col1 NUMBER,
    col2 VARCHAR2(255)
)And now I can insert each collection into that table as they're made by the functions (with a normal for loop). Now when I want the intersection, I'm trying to write a query like:
declare
  type test_record is record (
    col1 NUMBER,
    col2 VARCHAR2(255)
  type n_table is table of test_record;
  t3 n_table;
  total_collections NUMBER;
begin
    total_collections := 4; --this number I won't know at design time, its a counter to see how many collections I've created and inserted into the temporary table
    select col1, col2 bulk collect into t3 from TEMP_STORAGE
        where count(col1) = total_collections;
end;So I want to count the number of duplicates in col1 for a row, and if there are enough, then select that row. I know I can't use the count function like that, is there another way I can do this without creating a seperate table for each function, and doing an intersect on all of them?
Thanks again for your quick response! If you need any clarification about what I'm doing or asking, please ask.