Designing a appointment timetable (performance problem)

Argonist

I created a report for the appointment timetable. But I have a problem with performance. It takes about 7 sec.

I am using a table (7 columns) for report and in each cell also a table! It is too much?
Goal: appointment-timetable (weekly listing)

---

| monday | tuesday | ....
username | 12.00-13.00 |
| 14.00-16.00 |

---

username1 |
...

also daily listing and monthly listing

If a user has no appointment and should always appear into table.

I attached a sample!

Thanks

Clement Wong

With any performance issue, it's best to segment the various components and determine where the time is spent in each area.<br />
<br />
Without being able to run your report and briefly looking at your design, I can suggest the following:<br />
<br />
First, find out how long each of your queries for the Data Sets take to run on your database server and to return the results back.<br />
<br />
Second, since you are using MySQL, depending on your result set size, you may want to change the default value in your JDBC URL string. With the MySQL driver, the entire ResultSet is buffered in memory by default.<br />
<br />
<pre class='_prettyXprint _lang-auto _linenums:0'>jdbc:mysql://someDB.somewhere.com/someDatabase?useCursorFetch=true&defaultFetchSize=2000&zeroDateTimeBehavior
</pre>
Third and most likely a major factor is that the design has 3 Joined Data Sets, each with parameter to filter out unnecessary results. 4 out of the 5 (MySQL) Data Sets in your design appear to be returning a large amount of data (e.g. SELECT * ...)<br />
<br />
In general, for improved performance, it's best to push your joins and filters (and sorts) to your database server. It will help reduce the size of the query result set that the BIRT layer will need to handle.