Hi, there

I wonder if berkeley-db je provide bulk access methods when
sequentially accessing all the records in a large database? Does Cursor
just do it? Or any other suggestions to speedup the retrieval?

Thanks

Tong

Hello Tong,

No, JE does not currently have bulk access operations.

If the retreival is CPU bound, using non-transactional access or the
ReadCommitted isolation level may help improve performance.

If the retrieval is I/O bound, inserting the records initially in key
order (presorting if necessary) will speed up sequential retrieval.

I'm curious, do you need to retrieve in key order, or is order
unimportant? I ask this question because we are thinking about ways to
speed up retreival when order is not important.

For the quickest response on JE questions please use the bdbje mailing
list mentioned at the bottom of this page:

http://sleepycat.com/products/je.shtml

Mark

Mark,

The retrieval is I/O bound and the retrieval orde is not important.

Thanks

Tong

Mark,

The order doens't matter and the program is I/O bound.

Thanks

Tong

mark (AT) sleepycat (DOT) com wrote: