-- My original datafile
oracle@mutation#cat mydatafile
B1,xxx,bbb,bbb,bbb,bbb,bbb,bbb
H2,ccc,ccc,cc,cca
H2,ccc,ccc,cc,ccb
B1,yyyy,ddd,ddd,ddd,ddd,ddd,ddd
H2,ccc,ccc,ccc,ccc
B1,zzz,ddd,ddd,ddd,ddd,ddd,ddd
H2,ccc,ccc,ccc,ccd

--  SOMESCRIPT is the script i used
--
oracle@mutation#somescript
*********NewDatafileIs*****
H2,ccc,ccc,cc,cca,xxx
H2,ccc,ccc,cc,ccb,xxx
H2,ccc,ccc,ccc,ccc,yyyy
H2,ccc,ccc,ccc,ccd,zzz

-- source for somescript
--
oracle@mutation#cat somescript
cat mydatafile | while read thisline
do
 HD=`echo $thisline | awk -F',' '{print $1}'`
 if [ "$HD" = "B1" ]; then
 HT=`echo $thisline | awk -F',' '{print $2}'`
 else
   echo $thisline,$HT  >> /tmp/newdatafile
 fi
done
echo "*********NewDatafileIs*****"
cat /tmp/newdatafile
#-- use sqlldr to load this /tmp/newdatafile
#-- after loading rename/move/remove the newdatafile
rm /tmp/newdatafile
oracle@mutation#

CREATE TABLE dummy_b1 (val varchar2(255));

CREATE PACKAGE remember_me AS
    val VARCHAR2(255);
END;
/

CREATE TRIGGER dummy_b1_t1
BEFORE INSERT ON dummy_b1
FOR EACH ROW
BEGIN
    remember_me.val = :new.dummy;
END;
/

CREATE TRIGGER h2_table_t1
BEFORE INSERT ON h2_table
FOR EACH ROW
BEGIN
    :new.val := remember_me.val;
END;
/