![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
I need to dump Access MDB files to SQL Server 2005. No extract and transfer, just transfer the raw data. However, the mdb files can be reaching the limitation of 2G. Is there any benifit (especially performance) to use SSIS to load the MDB files? Or a simple SQL statement (with openquery to open the oledb Jet database, 'insert into '+@tablename + ' select * from openquery(....)) will be the same? |
#3
| |||
| |||
|
|
there is a big difference !!! SSIS use a bulk insert command which is far better then a simple insert command. you have to setup correctly the commit batch size. for example, if you have 1 million of rows, setup the batch size to 10 000 insert 10 000 rows, commit this batch, then load the next 10 000 rows.... instead-o creating 1 transaction for the million of rows, multiple smaller transaction are created. this reduce the time taken to load the data, reduce the log file size needed. so the answer is clear: use SSIS. "nick" <nick (AT) discussions (DOT) microsoft.com> wrote in message news:2705757C-0AE0-4830-969B-46EAD54A0BA7 (AT) microsoft (DOT) com... I need to dump Access MDB files to SQL Server 2005. No extract and transfer, just transfer the raw data. However, the mdb files can be reaching the limitation of 2G. Is there any benifit (especially performance) to use SSIS to load the MDB files? Or a simple SQL statement (with openquery to open the oledb Jet database, 'insert into '+@tablename + ' select * from openquery(....)) will be the same? |
#4
| |||
| |||
|
|
Yes, transaction batch is the different. However, how SSIS comparing with bulk insert? Anyway to implement a kind of bulk insert using TSQL that read ACCESS MDB files (via openrowset?)... "Jéjé" wrote: there is a big difference !!! SSIS use a bulk insert command which is far better then a simple insert command. you have to setup correctly the commit batch size. for example, if you have 1 million of rows, setup the batch size to 10 000 insert 10 000 rows, commit this batch, then load the next 10 000 rows.... instead-o creating 1 transaction for the million of rows, multiple smaller transaction are created. this reduce the time taken to load the data, reduce the log file size needed. so the answer is clear: use SSIS. "nick" <nick (AT) discussions (DOT) microsoft.com> wrote in message news:2705757C-0AE0-4830-969B-46EAD54A0BA7 (AT) microsoft (DOT) com... I need to dump Access MDB files to SQL Server 2005. No extract and transfer, just transfer the raw data. However, the mdb files can be reaching the limitation of 2G. Is there any benifit (especially performance) to use SSIS to load the MDB files? Or a simple SQL statement (with openquery to open the oledb Jet database, 'insert into '+@tablename + ' select * from openquery(....)) will be the same? |
![]() |
| Thread Tools | |
| Display Modes | |
| |