Relational databases are a mature technology, and they should be used to do as much simple work as possible.

For example, don't do the following in code, if they can be done in SQL instead:

• ordering (ORDER BY)
• filtering based on criteria (WHERE)
• joining tables (WHERE, JOIN) -- see the code tables topic for a possible exception to this rule
• summarizing (GROUP BY, COUNT, AVG, STDDEV)

Any corresponding task implemented by your own code would very likely

• be much less robust and efficient
• take longer to implement
• require more maintenance effort

To retain the order of a ResultSet as specified in ORDER BY, choose a collection which has a defined iteration order, such as an ArrayList or a LinkedHashMap .

Martin Fowler has published an interesting article related to this topic.

Implementing compareTo
Choosing the right Collection
Reduce database code duplication
Implement code tables

Yes   No   Undecided