I have been messing around with POI again a bit since with XSSF it can now support more than 65536 rows. Previously I was using ruby and the win32ole to solve my problems but it's slow so I'm back and looking for a better solution. Of course, when hitting such high numbers of rows I instantly ran into JVM heap issues. Then I found the BigGridDemo and it seemed to work nicely. I started trying to adapt it into a simple class I can use for outputting plain text data. No cell formats or fonts or anything fancy, just about ~100k rows of data. The problem is I have 3-4 sheets I need to create per workbook. I'm trying to figure out the inner workings of the BigGridDemo and what it's doing with the zip files and streams to get all this magic happening but I'm not having much luck. So my question is firstly - is the BigGridDemo method still the way to go for this? The last time I see mention of it is back in January and I'm not sure if there is a more recent 'best practice' for large data sets. If it is still relevant, can anyone give me an idea of how I can adapt this to multiple sheets? I've stepped through it quite a few times but I still cannot get a handle on what is happening in the substitute() function.
thanks Alex
