Topics in this issue:

• Wicket adoption in Macau
• How to set up and use auto-complete in Eclipse
• Improving database access performance from your program
  

Wicket adoption in Macau

Recently I learned that there are quite some organizations in Macau have adopted the excellent Apache Wicket framework for rapid web application development. These include a major higher education institute, two large government agencies, an small software start-up, in addition to CPTTM itself.

This information should be useful for you to persuade your management to adopt it, so that you can enjoy a much higher productivity (e.g., than Struts or JSF). If your organization is already using it, don't hesitate to let me know (I will not disclose the name of the organization unless you want me to) so that other poor programmers can benefit.

How to set up and use auto-complete in Eclipse

In Eclipse, if you like to type, say, the following line:

List l = ...;

You could type "Li" and then Alt+/ and it will suggest "java.util.List". This is called auto-complete or content assist. To enable this feature:

1. Choose Windows | Preferences | General | Keys.
2. Use "Word Completion" as the search key.
3. Remove its binding (by default, Alt+/ is bound to "Word Completion", not "Content Assist").
4. Use "Content Assist" as the search key.
5. Bind it to Alt+/.
   

Improving database access performance from your program

If your program is very slow when accessing the database, usually there are only two reasons: You are are fetching more information than you actually use or fetching the information in too many roundtrips. More specifically:

• Fetching more rows than needed. Will the SELECT statement match a lot of rows? Even if your code will only actually read a few of the rows, many JDBC drivers will fetch all matching rows from the database server before returning anything to your code. The solution is to use OFFSET and LIMIT (useful for paging). Another solution is to call, say, setFetchSize(50) on the query so that it will fetch at most 50 rows at one time. However, this is only a hint to the JDBC driver; it may or may not obey.
• Fetching more columns than needed. Are you fetching long columns such as those varchar(1024) by using "SELECT *"? If you don't need those columns, specify exactly those you really need such as:

SELECT ID, NAME from PRODUCTS

• Using many small queries. Are you selecting, say, all the products using one SELECT and then issue a SELECT to find all the parts for each product? If you have N products, you will issue N+1 queries. A much better way is to use a single query using join such as:

SELECT A.ID, A.NAME, B.NAME from PRODUCTS A JOIN PARTS B ON A.ID=B.PRODUCT_ID

Feedbacks