The short answer is that it is always faster to do a direct sql statement over the addnew/update. Depending on the cursor and other properties, ado has to manage quite a bit when using the addnew/update. When dealing with open recordsets using the addnew method, ado has to manage whether a current record was being edited and if so it call update - and when you use addnew, ado caches the changes locally then on update posts the changes to the database and changes properties for the recordset object ... so there is much more going on behind the scenes than just a straight sql statement that returns a closed recordset object. Use the connection's execute method for non-row returning sql commands - such as insert/update/delete.

Happy to help.

Here is my code:

set rs = server.CreateObject("adodb.recordset"
rs.CursorType = adOpenKeyset
rs.LockType = adLockOptimistic
rs.Open "tblClaimHistory", cnn, , , adCmdTable
lngOffset = 0
Do While lngOffset < lngFileSize
varChunk = leftB(RightB(varFDFContent, lngFileSize - lngOffset), iChunkSize)
rs.Fields("ClaimContent").AppendChunk varChunk
lngOffset = lngOffset + iChunkSize
Loop
rs.Update
rs.Close
set rs = nothing

----
I got timeout expired, and need to make this update goes faster.

So how do you replace:

rs.Fields("ClaimContent").AppendChunk varChunk

with update query?

Thanks.