Need help for understanding the behaviour of these 2 queries....

Question

Hi,
I need your help for understanding the behaviour of following two queries.
The requirement is to repeat the values of the column in a table random no of times.
Eg. A table xyz is like -
create table xyz as
select 'A' || rownum my_col
from all_objects
where rownum < 6;
my_col
A1
A2
A3
A4
A5
I want to repeat each of these values (A1, A2,...A5) multiple times - randomly decide. I have written the following query..
with x as (select my_col, trunc(dbms_random.value(1,6)) repeat from xyz),
y as (select level lvl from dual connect by level < 6)
select my_col, lvl
from x, y
where lvl <= repeat
order by my_col, lvl
It gives output like
my_col lvl
A1     1
A1     3
A1     5
A2     1
A2     3
A2     5
A3     1
A3     3
A3     5
A4     1
A4     3
A4     5
A5     1
A5     3
A5     5
Here in the output, I am not getting rows like
A1     2
A1     4
A2     2
A2     4
Also, it has generated the same set of records for all the values (A1, A2,...,A5).
Now, if I store the randomly-decided value in the table like ---
create table xyz as
select 'A' || rownum my_col, trunc(dbms_random.value(1,6)) repeat
from all_objects
where rownum < 6;
my_col repeat
A1     4
A2     1
A3     5
A4     2
A5     2
And then run the query,
with x as (select my_col, repeat from xyz),
y as (select level lvl from dual connect by level < 6)
select my_col, lvl
from x, y
where lvl <= repeat
order by my_col, lvl
I will get the output, exactly what I want ---
my_col ....lvl
A1     1
A1     2
A1     3
A1     4
A2     1
A3     1
A3     2
A3     3
A3     4
A3     5
A4     1
A4     2
A5     1
A5     2
Why the first approach do not generate such output?
How can I get such a result without storing the repeat values?

Answer 1

If I've understood your requirement, the below will achieve it:
SQL> create table test(test varchar2(10));
Table created.
SQL> insert into test values('&test');
Enter value for test: bob
old   1: insert into test values('&test')
new   1: insert into test values('bob')
1 row created.
SQL> insert into test values('&test');
Enter value for test: terry
old   1: insert into test values('&test')
new   1: insert into test values('terry')
1 row created.
SQL> insert into test values('&test');
Enter value for test: steve
old   1: insert into test values('&test')
new   1: insert into test values('steve')
1 row created.
SQL> insert into test values('&test');
Enter value for test: roger
old   1: insert into test values('&test')
new   1: insert into test values('roger')
1 row created.
SQL> commit;
Commit complete.
SQL> select lpad(test,(ceil(dbms_random.value*10))*length(test),test) from test;
LPAD(TEST,(CEIL(DBMS_RANDOM.VALUE*10))*LENGTH(TEST),TEST)
bobbobbobbobbobbobbobbobbobbob
terryterry
stevestevesteve
rogerrogerrogerrogerrogerrogerrogerrogerrogerYou can alter the value of 10 in the SQL if you want the potential for a higher number of names.
Andy