Speed issues in file parsing

Question

Hi everyone
I am a commercial developer who is relatively new to Java technology. My company has just finished the initial version of a servlet-based web program which is designed to interact with a MySQL database.
We are having a major performance issue, however. The servlet works by reading a raw HTML file on the server, and looking for field names (suggested by a marker). When it finds one, it replaces it with data from a resultset.
One of the most common replacements is for an autotext field. We have a problem however, in that one of our clients has 5000 records, and the program is taking nearly 30 seconds (on a Pentium 3 800mhz, 500meg ram, running nothing else except Apache Tomcat 4.03) to do the lines of code which send the resultset to the html. Specifically, it does this -
1. Reads the first entry in a vector.
2. Adds a line to a new vector which contains the (revised) html code.
3. Reads the next entry
and so on for 5000 records, and then, later,
1. Reads the first entry from the html (revised) vector
2. out.println's the line to the HTMLServerresponse/
3. Reads the next entry
for 5000 records.
In my thinking, reading a line of text and writing it to another vector should be faster than 10,000 records per 30 seconds.
Any suggestions?
Brendan

Answer 1

Hi.
1. Replace your Vectors with ArrayLists. This will stop mutexes having to be grabbed all the time
2. Make sure your file processing and output uses buffers (BufferedInputStream/ BufferedOutputStream/ BufferedReader/ BufferedWriter) - this may make a significant difference
But it sounds to me that what you are trying to do is treat the HTML as a template. You'd be better off using JSPs or Jakarta's Velocity stuff.
Regards,
Lance
Lance Walton - [email protected]
Team In A Box - Software without Tragedy
http://www.teaminabox.co.uk