ODBC for WonderWare Historian

Hi All,

First post here.

I am having some issues extracting large amounts of tag history from WonderWare into another historian.
Unfortunately OPC isn't performing the best, so my only other options is through ODBC.
(For more details on OPC.. I am using HDA so I know the performance won't be the best as it's constantly pulling values from the last time stamp that was collected.. Also I notice that the OPC Server for inSQL just connects to Microsoft SQL Server and executes queries... and my whole intent is to try and eliminate MS SQL Server from the equation)

Does anybody know if there is a native ODBC driver for WonderWare inSQL (Historian), and by this i mean an ODBC connection that does not connect to inSQL through Microsoft SQL Server?

If there were an ODBC that connected directly to inSQL rather than using 'linked server' through SQL Server, I would assume the performance may increase.

If you dont know, but have experience with polling large amounts of data from WonderWare historian, any information would be appreciated.

Thanks!

RDS
OSISoft PI System Administrator

It's possible I am misunderstanding the issue, but here is how I see it.

OPC is the normal method used to put data into InSQL historian.

InSQL stores it's configuration data in SQL tables, but not the actual historical data. Historical data is stored in a proprietary format. InSQL provides you with "views" that respond to SQL queries just like a SQL view or table would. My experience with accessing the historical data using SQL queries of the provided views is that it handles large amounts of data quickly. This is the normal access for reporting.

I can't be 100% sure at this point, but I don't think SQL server is involved in these transactions. SQL server is definitely involved in configuration. I guess I have to ask... Why are you trying to eliminate SQL server when it is a required component of InSQL?

Caveat: This is all based on experience that is a few years old and things change.

Thanks for the response mellis.
I do not have too much experience with InSQL, but yes you are correct.
They store the actual values in history blocks (flat files) that are basically attached to the SQL server via views.
OPC is used to put data INTO InSQL historian... but on the other end, I am trying to pull all the data OUT of historian.
Right now I am using Matrikon OPC for inSQL, but the performance is not too great. I am only polling 7500 tags for now, but will eventually be polling over 250k tags. I was just hoping that there was a better, or faster/more effecient way in retrieving such large amounts of data.

Update:
I just noticed your followup question. The only reason i am trying to eliminate sql server from the equation is just the hope that there is a way to access the data more effeciently. Right now the performance of pulling such large amounts of data from those views are providing a bit slow. I am currently restricted to only using HDA rather than DA in order to not miss any data, but we are also in the beginning stages of testing an RDBMS interface through ODBC for retrieving values. (Although the RDBMS interface through ODBC will essentially be executing the same queries as OPC for inSQL is, hence the search for a more effecient way to retrieve data).

thanks again.

RDS