问题
I'm using NPOI to deal with Excel files. Here's how I'm reading files:
using (FileStream stream = File.OpenRead(excelFilePath))
{
IWorkbook wb = WorkbookFactory.Create(stream);
...
}
However, for any XLSX file larger than a few megabytes, it causes memory usage to shot up to about 1GB and eventually throw an OOM exception.
Doing some research, I've found out that, strangely, loading a workbook from a File rather than a Stream results in less memory consumption by POI. The closest C# equivalent to the provided Java examples I've come up with to use Files is the following:
OPCPackage pkg = OPCPackage.Open(new FileInfo(excelFilePath));
XSSFWorkbook wb = new XSSFWorkbook(pkg);
But it seems to use the same underlying implementation since the memory usage is still the same and causes OutOfMemory exceptions.
Does NPOI have anything built-in for handling large XLSX files?
Suggestions on alternative libraries that can handle both XLS and XLSX files are also welcome.
回答1:
It seems XLSX support is rather new in NPOI and it simply can't handle large files yet.
After trying a few libraries, EPPlus was able to process the large XLSX file without a hitch, so I finally settled on having two libraries for reading in Excel files, NPOI for XLS and EPPlus for XLSX.
回答2:
As a suggestion for alternative libraries, a good one is Apache POI. I've used it extensively for both XLSX and XLS files and it does the job well. Here's a gist to do a quick test on your file.
The only format Apache POI doesn't cover is the old format XML files for which Xelem can be used instead.
来源：https://stackoverflow.com/questions/33670705/outofmemoryexception-reading-20mb-xlsx-file