Approach for coding a business req

Question

Hi ,
I have a business rule for populating the target table from few source tables. Please have a look at the rule.What should be the approach to handle such business rules .The number of rows in the tables are less than a million .I cant really provide a sample data ,apologies .But still would need a rough approach that one should follow.
if nvl(source_table_1.col1,source_table_1.col2)
then
   if nvl(source_table_1.col1,souce_table_1.col2)=source_table_2.col1
     then
      target_table_1.col1=source_table2.col1
      target_table_1.col2=source_table2.col2
      target_table_1.col3=source_table2.col3
      target_table_1.col4=source_table2.col4
      target_table_1.col5=source_table3.col1
      target_table_1.col6=source_table5.col1
  else if
     source_table_1.col8=source_table_6.col1 where nvl(source_table_1.col1,source_table_1.col2)=
                                         source_table_7.col1
        then
      target_table_1.col1=source_table2.col1
      target_table_1.col2=source_table2.col2
      target_table_1.col3=source_table2.col3
      target_table_1.col4=source_table2.col4
      target_table_1.col5=source_table3.col1
      target_table_1.col6=source_table5.col1
   else
   null
   end
else
null
end;

Answer 1

Hi,
A single MERGE statement should do what you want; no PL/SQL is required.
It looks like you have 2 somewhat different ways of getting the new data, one preferred to the other. In the MERGE ... USING clause, do a Top-N Query, where you get the data both ways. That will give you sets of new data, with up to 2 rows in each set. Number the rows in each set in preferential order, then pick the rows numbered 1.
Rahul  K wrote:
... I cant really provide a sample data ,apologies .But still would need a rough approach that one should follow.If you post a little sample data, then you can get more than just a rough approach.