I can remove records from the beginning, middle, or end, and it won't load until the file is smaller than 2MB. And yes, you are right, 2MB isn't very many records (just over 2000, in my case). Remember, that the records have a length of the last column being "CX", which is very long.
As far as malformed files, that logic would mean that I wouldn't be able to use cfspreadsheet on any of the files. The tag works great on the files less than 2MB, that I downloaded directly from the portal.
As it stands right now, I have to chop up the files into 1000-record "bite site" smaller files. After doing that, cfspreadsheet breezes right through them.
When I first saw this problem a few months back, I poured hours into searching it up and down on Google and found very little. What I did find was mainly about problems and bugs in the Java code (nothing having to do with CF, just the Java function(s) that CFSpreadsheet uses apparently). All of these hits on Google being years old.
I do not know what is causing the problem. But having read much bigger files, with more columns, I am not yet convinced it is solely a size problem.
- Can you read any files over 2MB or ones with a similar number of rows and columns?
- Any odd characters or charsets involved?
- Any idea what program produced the files? Either from viewing the metadata or the headers?
When I first saw this problem a few months back, I poured
hours into searching it up and down on Google and found very
little.
Did you ever try running the files through POI's BiffViewer?
AdChoices