Creating and displaying Varrays in Apex 4.1

Question

Hi,
I am creating quite a sophisticated database, which has to support relations between different tables. In order to implement that, I've decided to use user-defined datatype - a varray which consists of 100 integers (and only of that) that will contain ID's to different rows in different tables. Since it is a user-defined datatype I have problems with using it in Application Express :)
First question: Is there any way to make this simple datatype 'readable' for Apex? In each application I create it shows that it is 'an unsupported datatype'. Maybe I should define something more than just varray construction - the way Apex should read and edit it etc ?
Second question: How can I implement a function that will read the IDs in the varray and transform them into names to which that IDs correspond? I would like an application to show full name of an organization (ex. "ABC Corporation") instead of its ID. I tried to put some part of code that takes up the name of an organisation in a row with given ID into +"Source"+ form in the +"Edit page item"+, however Apex doesn't seem to get the fact that the data is in different table. Any ideas? :)
I will be grateful for any help since it is an urgent case :]

Answer 1

KamilGorski wrote:
I would happily do that if only I had more time to study SQL and learn it properly :) Unfortunately, our start-up company has no time and money at the moment to allow me to do so.Then isn't using technologies that no one knows rather a strange decision?
But coming back to your solution - it still works only if user inputs only one product quality. Let's say that user has chosen 3 qualities and the 'SELECTED_QUALITIES' collection looks like that:
n001 = 1,
n002 = 3,
n003 = 5
And since the SELECT query you have created compares pq.quality_id only to n001, it returns all the products that have the quality no 1 - not all the products that have all three selected qualities - 1, 3 and 5. Of course, we can change the 'EXISTS' condition and add more 'OR' conditions like that:
where pq.quality_id = q.n001
or pq.quality_id = q.n002
or pq.quality_id = q.n003But it has few flaws - first we assume that we know the number of qualities user has selected - 3 (and we don't know that), You've misunderstood. SQL is row based. To handle multiple values, we create more rows, not additional columns. In your preferred terms, the result of any query is an <i>m</i>&times;<i>n</i> array of columns and rows, where the number of columns <i>m</i> is fixed, and the number of rows <i>n</i> varies according to the query predicates.
It is not necessary to know the number of selected qualities, simply create a row in the collection for each selected quality to give a set of selected qualities.
The SELECTED_QUALITIES collection should look like:
N001
   1
   3
   5
secondly the query will return all the products that have one of the three selected qualities (and we want products that have all three qualities). That wasn't really clear from the information previously given, but it's certainly possible. With <tt>product_qualities(product_id, quality_id)</tt> as a primary/unique key, and with no duplicates in the selected qualities, a solution is the set of all products with the selected qualities, where the number of qualities matched for each product equals the number of qualities selected, as in this example:
SQL> create table products (
  2      product_id    integer       primary key
  3    , product_name  varchar2(30)  not null);
Table created.
SQL> create table qualities (
  2      quality_id    integer       primary key
  3    , quality_name  varchar2(30)  not null);
Table created.
SQL> create table product_qualities (
  2        product_id  integer   not null references products
  3      , quality_id  integer   not null references qualities,
  4      constraint product_qualities_pk primary key (
  5            product_id
  6          , quality_id))
  7    organization index;
Table created.
SQL> create index product_qualities_ix2 on product_qualities (
  2      quality_id
  3    , product_id);
Index created.
SQL> insert all
  2    into products (product_id, product_name) values (1, 'widget')
  3    into products (product_id, product_name) values (2, 'thingummy')
  4    into products (product_id, product_name) values (3, 'whatsit')
  5    into products (product_id, product_name) values (4, 'gizmo')
  6    into products (product_id, product_name) values (5, 'gadget')
  7    into products (product_id, product_name) values (6, 'contraption')
  8  select * from dual;
6 rows created.
SQL> insert all
  2    into qualities (quality_id, quality_name) values (1, 'green')
  3    into qualities (quality_id, quality_name) values (2, 'silver')
  4    into qualities (quality_id, quality_name) values (3, 'shiny')
  5    into qualities (quality_id, quality_name) values (4, 'furry')
  6    into qualities (quality_id, quality_name) values (5, 'digital')
  7    into qualities (quality_id, quality_name) values (6, 'hd')
  8    into qualities (quality_id, quality_name) values (7, 'wireless')
  9  select * from dual;
7 rows created.
SQL> insert all
  2    into product_qualities (product_id, quality_id) values (1, 1)
  3    into product_qualities (product_id, quality_id) values (1, 3)
  4    into product_qualities (product_id, quality_id) values (2, 2)
  5    into product_qualities (product_id, quality_id) values (2, 4)
  6    into product_qualities (product_id, quality_id) values (3, 1)
  7    into product_qualities (product_id, quality_id) values (3, 3)
  8    into product_qualities (product_id, quality_id) values (3, 5)
  9    into product_qualities (product_id, quality_id) values (4, 2)
10    into product_qualities (product_id, quality_id) values (4, 4)
11    into product_qualities (product_id, quality_id) values (4, 6)
12    into product_qualities (product_id, quality_id) values (5, 2)
13    into product_qualities (product_id, quality_id) values (5, 3)
14    into product_qualities (product_id, quality_id) values (5, 5)
15    into product_qualities (product_id, quality_id) values (6, 1)
16    into product_qualities (product_id, quality_id) values (6, 3)
17    into product_qualities (product_id, quality_id) values (6, 5)
18    into product_qualities (product_id, quality_id) values (6, 7)
19  select * from dual;
17 rows created.
SQL> commit;
Commit complete.For the purposes of creating a quick and simple example outside of APEX, I'm using a temporary table instead of the SELECTED_QUALITIES APEX collection. For various reasons collections work better in APEX than GTTs and are the preferred approach for temporary storage.
SQL> create global temporary table selected_qualities (
  2    n001 integer)
  3  on commit delete rows;
Table created.With one quality selected, we get all the products having that quality:
SQL> insert into selected_qualities (n001) values (1);
1 row created.
SQL> insert into selected_qualities (n001) values (1);
1 row created.
SQL> select
  2            p.product_id
  3          , p.product_name
  4  from
  5            products p
  6           join product_qualities pq
  7             on p.product_id = pq.product_id
  8           join selected_qualities sq
  9             on pq.quality_id = sq.n001
10  group by
11            p.product_id
12          , p.product_name
13  having
14             count(*) = (select count(*) from selected_qualities);
PRODUCT_ID PRODUCT_NAME
      1 widget
      6 contraption
      3 whatsitThese products all have the next quality added, so continue to be returned:
SQL> insert into selected_qualities (n001) values (3);
1 row created.
SQL> select
  2            p.product_id
  3          , p.product_name
  4  from
  5            products p
  6           join product_qualities pq
  7             on p.product_id = pq.product_id
  8           join selected_qualities sq
  9             on pq.quality_id = sq.n001
10  group by
11            p.product_id
12          , p.product_name
13  having
14             count(*) = (select count(*) from selected_qualities);
PRODUCT_ID PRODUCT_NAME
      1 widget
      6 contraption
      3 whatsitThen as additional qualities are selected, the result set is progressively reduced:
SQL> insert into selected_qualities (n001) values (5);
1 row created.
SQL> select
  2            p.product_id
  3          , p.product_name
  4  from
  5            products p
  6           join product_qualities pq
  7             on p.product_id = pq.product_id
  8           join selected_qualities sq
  9             on pq.quality_id = sq.n001
10  group by
11            p.product_id
12          , p.product_name
13  having
14             count(*) = (select count(*) from selected_qualities);
PRODUCT_ID PRODUCT_NAME
      6 contraption
      3 whatsit
SQL> insert into selected_qualities (n001) values (7);
1 row created.
SQL> select
  2            p.product_id
  3          , p.product_name
  4  from
  5            products p
  6           join product_qualities pq
  7             on p.product_id = pq.product_id
  8           join selected_qualities sq
  9             on pq.quality_id = sq.n001
10  group by
11            p.product_id
12          , p.product_name
13  having
14             count(*) = (select count(*) from selected_qualities);
PRODUCT_ID PRODUCT_NAME
      6 contraption