Insert records from a single table to two related tables

Question

DB - 10G
OS - XP
We have many thousands of comma delimited records that we want to insert into a normalised table structure.
I have created a test dataset that can be found at the end of this post.
I have csv records that look like this;
donald, huey
donald, dewey
donald, louie
And I want the data to be inserted into two separate table like this;
table named PARENTS
pk parent_name
1  donald
Table named CHILDRENS
pk fk child_name
1  1 huey
1  2 dewey
1  3 louie
I constructed an insert statement that looks like this;
INSERT ALL
  INTO parents (parent_id, parent_name) VALUES (parents_seq.nextval, parent_name)
  INTO children (children_id, parent_id, child_name) VALUES(children_seq.nextval, parents_seq.nextval, child_name )
SELECT parent_name, child_name
  FROM sources;
And this resulted in the following;
Table named PARENTS
pk child_name
1  DONALD
2  DONALD
3  DONALD
Table named PARENTS
pk fk child_name
1  1  HUEY
2  2  DEWEY
3  3  LOUIE
Would anyone have any ideas on how I could accomplish this task?
This is some example data;
[code]
DROP SEQUENCE parents_seq;
DROP SEQUENCE sources_seq;
DROP SEQUENCE children_seq;
DROP TABLE sources;
DROP TABLE children;
DROP TABLE parents;
CREATE SEQUENCE PARENTS_SEQ MINVALUE 1 MAXVALUE 1000000000000000000000000000 INCREMENT BY 1 START WITH 1 CACHE 20 NOORDER NOCYCLE;
CREATE TABLE PARENTS
( PARENT_ID      NUMBER(8) NOT NULL
, PARENT_NAME    VARCHAR2 (50 CHAR) NOT NULL
, CONSTRAINT     PARENTS_PK PRIMARY KEY (PARENT_ID)
create or replace
TRIGGER PARENTS_BI
BEFORE
  INSERT OR UPDATE ON PARENTS
  FOR EACH ROW BEGIN
  IF INSERTING THEN
    IF :NEW.PARENT_ID IS NULL THEN
      SELECT PARENTS_SEQ.nextval INTO :NEW.PARENT_ID FROM dual;
    END IF;
  END IF;
END;
CREATE SEQUENCE SOURCES_SEQ MINVALUE 1 MAXVALUE 1000000000000000000000000000 INCREMENT BY 1 START WITH 1 CACHE 20 NOORDER NOCYCLE;
CREATE TABLE SOURCES
( SOURCE_ID      NUMBER(8) NOT NULL
, PARENT_NAME    VARCHAR2 (50 CHAR) NOT NULL
, CHILD_NAME    VARCHAR2 (50 CHAR) NOT NULL
, CONSTRAINT     SOURCES_PK PRIMARY KEY (SOURCE_ID)
create or replace
TRIGGER SOURCES_BI
BEFORE
  INSERT OR UPDATE ON SOURCES
  FOR EACH ROW BEGIN
  IF INSERTING THEN
    IF :NEW.SOURCE_ID IS NULL THEN
      SELECT SOURCES_SEQ.nextval INTO :NEW.SOURCE_ID FROM dual;
    END IF;
  END IF;
END;
INSERT INTO SOURCES (PARENT_NAME, CHILD_NAME) VALUES ('DONALD', 'HUEY');
INSERT INTO SOURCES (PARENT_NAME, CHILD_NAME) VALUES ('DONALD', 'DEWEY');
INSERT INTO SOURCES (PARENT_NAME, CHILD_NAME) VALUES ('DONALD', 'LOUIE');
Commit;
CREATE SEQUENCE CHILDREN_SEQ MINVALUE 1 MAXVALUE 1000000000000000000000000000 INCREMENT BY 1 START WITH 1 CACHE 20 NOORDER NOCYCLE;
PROMPT *** CREATE TABLE ***
CREATE TABLE CHILDREN
( CHILDREN_ID    NUMBER NOT NULL
, PARENT_ID      NUMBER NOT NULL
, CHILD_NAME     VARCHAR2 (50 CHAR) NOT NULL
, CONSTRAINT     CHILDREN_PK PRIMARY KEY (CHILDREN_ID)
create or replace
TRIGGER CHILDREN_BI
BEFORE
  INSERT OR UPDATE ON CHILDREN
  FOR EACH ROW BEGIN
  IF INSERTING THEN
    IF :NEW.CHILDREN_ID IS NULL THEN
      SELECT CHILDREN_SEQ.nextval INTO :NEW.CHILDREN_ID FROM dual;
    END IF;
  END IF;
END;
[code]

Answer 1

Looks like this is one way of achieving it:
insert into parents (parent_name) select distinct parent_name from sources;
select *
  from parents;
PARENT_ID              PARENT_NAME                                       
1                      DONALD
insert into children (parent_id, child_name)
select p.parent_id, s.child_name
  from sources s
  join parents p
    on (s.parent_name = p.parent_name);
select *
  from children;
CHILDREN_ID            PARENT_ID              CHILD_NAME                                        
1                      1                      HUEY                                              
2                      1                      DEWEY                                             
3                      1                      LOUIE