Hi Doug,
As opposed to what other mechanism?
DataReader is used internally by DataAdapter, so I really don't see any other
fully-managed choice but to directly or indirectly use a DataReader.
Also, the latency depends on the amount of data in each row. If each of the
10000 rows are reasonably small in size the operation should be quick on a
LAN. Figure that if each row has a maximum size of 100 B then the total size
of the 10000 rows of data won't exceed 1 MB (+ protocol control bits). On a
10Mb/s LAN the total latency would be ~1 second.
I agree with Marc that your choice of queried data is probably more important
than how you get it into your application. After all, decreasing the size of
a returned row, 10000 times, is going to help out more than switching the
mechanism that downloads them.
If you are still seeing performance problems, then mock up a program to
compare the performance with a DataReader vs. a DataReader used by a
DataAdapter vs. any other mechanism you have in mind. Choose the quickest of
them all :)
--
Dave Sexton
"Doug" <dn******@dtgnet.comwrote in message
news:11*********************@e3g2000cwe.googlegrou ps.com...
If I am grabbing a lot of rows (say 10,000 from SQL) would I be helping
or hurting the network by using a DataReader? I would think helping
because it's really only returning 1 row at a time, but I'm not sure.
Are there good articles out there that explain how the DataReader does
this as well as any that discuss performance problems/benefits?