An online comparison of POI and JXL

Poi is a subproject of the Apache company that provides a set of Java APIs to manipulate Windows documents.
Java Excel, commonly known as JXL, is an open source project through which Java developers can read the contents of an Excel file, create a new Excel file, and update an existing Excel file.
Using this API, a non-Windows operating system can also handle Excel data tables through a pure Java application. Because it is written in Java, we can use JSP, servlet to invoke the API to access the Excel data table in the Web application.

The difference between the two, mainly on the JVM virtual machine memory consumption situation.
Data volume 3,000 data, 60 columns per bar. JVM Virtual machine memory size 64M.
Use POI: Run to 2,800 or so to report memory overflow.
Use the jxl:3000 strip all out, and the memory has 21M of space.
It is conceivable that the memory consumption of the gap is still quite large.
Maybe it's because JXL is doing a pretty good job of recycling resources.
As to the efficiency of both, I think this is also based on the large data volume, the small amount of data is not very different, it is not difficult to be detected. But with a large amount of data, poi consumes much more of the JVM memory than JXL.
But the JXL is relatively weak compared to the functionality provided. So if you want to implement a more complex scenario, consider using POI, but if you just want to generate some big data, consider using JXL, or CSV is a good choice.
However, CSV is not really excel.

An online comparison of POI and JXL