Hi all,
We’re using poi & poi-ooxml 3.14 generating tens of thousands of rows to be
downloaded.
We have a problem that jvm keeps old gc 5 times/min but nothing could be
collected. What’s more, when we looked into the jvmdump, it has 2G unreachable
objects closed related with poi. We
Without more detail, my guess would be that your JVM has 2GB of XML nodes.
Nothing is flushed to disk with XSSF until you call Workbook.write, and
nothing is cleared until Workbook.close.
If you only need to create a workbook, the SXSSF streaming API is the way
to go.
On Sep 4, 2016 7:45 PM, "Hon
Hi Javen,
Per my understanding, before call Workbook.write, the XML nodes shouldn’t be
unreachable, the workbook should have strong reference to the nodes. Is that
right?
Seems the attached picture isn’t working. Here’s our code. The output stream is
from servlet response. The valueList(rows)
Did you try os.close() and wb.close() after your flush?
On Sep 4, 2016 20:50, "Hongdi Ren" wrote:
> Hi Javen,
>
> Per my understanding, before call Workbook.write, the XML nodes shouldn’t
> be unreachable, the workbook should have strong reference to the nodes. Is
> that right?
>
> Seems the att
By the way, if your production code really is as simple as the example you
provided, the only change you need to make to get this to work for
streaming is
Workbook wb = new SXSSFWorkbook(1);
Keep in mind that SXSSFWorkbooks don't use a shared string table--instead
they embed the string literal as
Thanks very much for the suggestions. I’ll try them
On 9/5/16, 1:08 PM, "Javen O'Neal" wrote:
By the way, if your production code really is as simple as the example you
provided, the only change you need to make to get this to work for
streaming is
Workbook wb = new SXSSFWo
