Connect by/Analytical problem

Question

I'm using Oracle 9i and have problem like below.
I have a table like this: header_part, component_part, qty
where component_part is child of header_part and qty is number of component parts.
One component part could consist of zero or more other component parts, i.e. a tree structure.
The same component can exist in one or many branches of the same tree.
Example:
header_part, component_part, qty
a b 2
a c 1
b d 1
b e 2
c d 2
c f 1
d g 1
Result:
component_part, tot_qty
b 2
c 1
d 4
e 4
f 1
g 4
What I like to do is to calculate the total qty for each component part for a given header part.
Is this possible in a singel query? How could I do this?

Answer 1

Why not just CONNECT BY and SUM?
Oracle Database 10g Express Edition Release 10.2.0.1.0 - Production
SQL> VARIABLE header_part VARCHAR2 (1);
SQL> EXEC :header_part := 'a';
PL/SQL procedure successfully completed.
SQL> SELECT :header_part, component_part, qty
  2  FROM    table_name
  3  START WITH header_part = :header_part
  4  CONNECT BY header_part = PRIOR component_part;
:HEADER_PART                     C        QTY
a                                b          2
a                                d          1
a                                g          1
a                                e          2
a                                c          1
a                                d          2
a                                g          1
a                                f          1
8 rows selected.
SQL> SELECT  :header_part, component_part, SUM (qty) total_qty
  2  FROM    table_name
  3  START WITH header_part = :header_part
  4  CONNECT BY header_part = PRIOR component_part
  5  GROUP BY   :header_part, component_part;
:HEADER_PART                     C  TOTAL_QTY
a                                f          1
a                                d          3
a                                g          2
a                                b          2
a                                e          2
a                                c          1
6 rows selected.
SQL>