May 14, 2010 at 5:37 pm
I agree with Elliott, you can't get away from transaction logging. As regards processing everything in memory, even SSIS is capable of doing that, only when you think the performance is suffering should you move that processing out of the Data flow by using alternatives like Execute SQL Task.
Amol Naik
May 14, 2010 at 5:47 pm
I consider SSIS a dedicated ETL tool. It does things that you can't easily do in SQL Server otherwise, and even if it doesn't do everything that another tool does, likley it does most of what you need.
And you can program the rest if you need it. It comes with your SQL Server license, might as well use it.
May 15, 2010 at 4:27 pm
I'm an old SQL Server guy, so I agree with you guys. It's just that we've had some proof of concept tests, with me advocating SSIS. For one thing, Informatica is substantially faster than SSIS 2005 (2008 might be faster than 2005). In addition, we couldn't find an adequate to export data from SQL Server 2005 to Oracle and there was no driver to properly connect to Vertica.
Back to the log file, I am trying to take a group of .csv files as an input and create a different set of .txt files as an output. In a case like that, Informatica might be a better tool, although I would prefer SSIS.
May 18, 2010 at 10:52 am
imani_technology (5/15/2010)
I'm an old SQL Server guy, so I agree with you guys. It's just that we've had some proof of concept tests, with me advocating SSIS. For one thing, Informatica is substantially faster than SSIS 2005 (2008 might be faster than 2005). In addition, we couldn't find an adequate to export data from SQL Server 2005 to Oracle and there was no driver to properly connect to Vertica.Back to the log file, I am trying to take a group of .csv files as an input and create a different set of .txt files as an output. In a case like that, Informatica might be a better tool, although I would prefer SSIS.
There are options available on the market for working with Oracle. If you can use third-party solutions, check CozyRoc Oracle Destination component. It is 30x faster compared to the standard OLE DB Destination. For Vertica check CozyRoc ODBC Destination component. If Vertica's ODBC driver supports the bulk-load API, you can expect substantial performance gains.
May 18, 2010 at 11:10 am
We use Data Integrator because it does something we need that SSIS does not do. We get packed decimal from our data supplier and Integrator can read the copylib and create a table from that. For anything else, we could use SSIS.
May 18, 2010 at 11:29 am
Steve Cullen (5/18/2010)
We use Data Integrator because it does something we need that SSIS does not do. We get packed decimal from our data supplier and Integrator can read the copylib and create a table from that. For anything else, we could use SSIS.
You can use Lysine[/url] for SSIS. It allows you to process packed decimals.
May 18, 2010 at 3:16 pm
imani_technology (5/15/2010)
I'm an old SQL Server guy, so I agree with you guys. It's just that we've had some proof of concept tests, with me advocating SSIS. For one thing, Informatica is substantially faster than SSIS 2005 (2008 might be faster than 2005). In addition, we couldn't find an adequate to export data from SQL Server 2005 to Oracle and there was no driver to properly connect to Vertica.Back to the log file, I am trying to take a group of .csv files as an input and create a different set of .txt files as an output. In a case like that, Informatica might be a better tool, although I would prefer SSIS.
I have just received feedback Vertica supports the bulk-load ODBC API. So the CozyRoc ODBC Destination might come handy.
July 28, 2023 at 2:06 pm
This was removed by the editor as SPAM
July 28, 2023 at 7:29 pm
Viewing 10 posts - 16 through 24 (of 24 total)
You must be logged in to reply to this topic. Login to reply