none
DQS vs MDS

    問題

  • Hi,

    I need to know if it exists any documentations about the matching between the new Data Quality Services and MDS. I think that DQS represents a specific task of the MDM process, isn't it?

    Thanks

    2012年3月5日 上午 11:32

所有回覆

  • The Matching option in MDS (only via the Excel Add-in for MDS that comes new in SQL Server 2012) talks to DQS and runs the matching algorithms to find potential duplicates.

    Match Data before Loading
    Before adding more data to MDS, you can now confirm that you aren’t adding duplicate records. The MDS Add-in for Excel uses SQL Server Data Quality Services (DQS) to compare two sources of data: the data from MDS and the data from another system or spreadsheet. DQS provides suggestions for updating your data, along with the percent of confidence that the changes are correct.
    Learn how to use the DQS component of the Add-in for Excel (XPS document) 

    DQS is the tool to do cleansing of data (finding mistakes) and matching (finding duplicates). These activities can be done inside DQS, matching can be done from MDS Excel Add-in, and cleansing can be done from SSIS dataflow task with the cleansing component there.

    Videos for more info :http://social.msdn.microsoft.com/Forums/en-US/sqldataqualityservices/thread/9513ecee-5d8b-482a-96ac-e6cfb7bff821


    Didn't get enough help here? Submit a case with the Microsoft Customer Support team for deeper investigation - http://support.microsoft.com/select/default.aspx?target=assistance

    2012年3月5日 下午 03:17
  • Hi Jason,

    I know that DQS is a tool for the data cleansing and the data matching, but I want to understand better the relationship with MDS. In a master data management process, do I need to use first DQS and then MDS? Does MDS contains some DQS features?

    Thanks

    2012年3月5日 下午 06:27
  • You can use the features totally independently if you want. They are not required in a certain order.

    You can certainly use the DQS+MDS features together. For example, you can do cleansing of data in DQS (maybe even in SSIS) before doing the MDS staging process to load the good data values into MDS.

    Or you can do the reverse - by pointing the DQS Cleansing project source to one of the MDS subscription views to find potential mistakes that exist in the MDS entities, and have the Data Steward export the corrected results for import back into MDS via the staging process, or manually make the equivalent corrections in MDS.

    Once data is in MDS, you can do the matching (Excel add-in for MDS ribbon has a button to find Matches) to find potential duplicates in your MDS Entity. Or you can point the DQS Matching project to a table from MDS Subscription Views to run the matching in DQS Client.

    At some point you do have to pull the data and the reference knowledge in and out between MDS and DQS, so there is some duplication of data values. What I mean is that MDS and DQS operate on separate catalogs to store data. The DQS KB holds the known good values. The MDS entity holds it own copy of the entity values.  MDS Staging process and the subscription views are the systematic interface points between the two sets of knowledge. Similarly DQS Data can be imported and exported to Excel, and MDS works well in Excel in SQL 2012 (the new Excel Add-in for MDS), so you can use the Excel worksheets as an an interactive interface point between the systems as well.

    I will ask around if we have any public topic documenting the interaction between systems for you to read more or see a demo.

    Thanks, Jason


    Didn't get enough help here? Submit a case with the Microsoft Customer Support team for deeper investigation - http://support.microsoft.com/select/default.aspx?target=assistance

    2012年3月5日 下午 08:05
  • Hi Jason, thanks for your interesting reply. I understand that DQS is a new solution and it could be useful to appreciate it better. I think (right?) that MDS is a solution for master data management issues and DQS is a tool for the data quality issues. With these different goals, DQS could be considered with a tool to aim to solve master data management issues and not a alternative solution otherwise, why using MDS?

    You are described some use scenarios, but I think it could be useful (and not only for me but for the entire forum) to list better the typical functional cases about using MDS itself, DQS itself and using MDS and DQS together. F.e.:

    - First case for using DQS and MDS together: first executing data cleansing with DQS, second loading cleansed data in the MDS staging tables and then matching data inside MDS ...

    It could be useful to have an article about this subject, describing more the functional point of view that the technologies or the components (Excel add-ins).

    I think it is useful to know the reasons to use MDS and the reasons to use DQS, otherwise one of these technologies could be viewed as a duplicate of the other one and so, why using both?

    Many many thanks!

    2012年3月6日 上午 07:59
  • There are plans to present more on this in the near future on blogs at perhaps at TechEd conference, so I won't steal any thunder by elaborating too much now.

    They are not duplicate - DQS checks semantic, quality, spelling, and accuracy of data. MDS is a master data management tool for organizing and entering and managing the data between multiple systems. They have bits of overlap (they both have some validation rules), but also a lot of distinctions as well.


    Didn't get enough help here? Submit a case with the Microsoft Customer Support team for deeper investigation - http://support.microsoft.com/select/default.aspx?target=assistance

    • 已提議為解答 Paras Doshi 2012年12月30日 上午 01:22
    2012年3月6日 下午 02:21
  • Hi Jason, it's a pity it hasn't major details about MDS and DQS scenarios. I hope to see as soon as possible a such subject treated on a blog.

    I think that I will not have other replies to my issue.

    Bye

    2012年3月6日 下午 08:59
  • pscorca, I'm glad someone asked this question the way you did.  Your point is exactly where I am in my evaluation of the use of QDS and/or MDS as a integral tool for this BI solution I am beginning to develop.  I hope to hear more detail from the SQL Team on this very question.

    Jason, I would be grateful if you would post some of the links to the information you infere will be coming in the near future to this post.


    SOTATS, Inc.

    2012年8月5日 下午 08:03
  • Hello,

    You might find the following white paper useful: Cleanse and Match Master Data by Using EIM.

    There is a section in this white paper that provides a comparison between DQS and MDS.

    Thanks,
    Vivek
    SQL Server Documentation

    • 已提議為解答 Paras Doshi 2012年12月30日 上午 01:21
    2012年11月2日 上午 10:16
  • Hi Kumar thanks for your reply.

    I hope to find a comparison to a business level and not to a technical level.

    I think DQS onto to underlying level respect to MDS. To introduce MDS into the market the idea could be to think MDS as the tool to create and mantain master data following certain business rules, but the candidate data to enter into the master data store must have a good quality guaranteed by DQS. So, MDS could be addressed to the company management and DQS could be seen as a tool in charge to the IT staff.

    Thanks

    2012年11月30日 上午 11:12
  • Hello pscorca,

    The white paper that I mentioned earlier does cover some use cases as well that highlight how and when you can use MDS and DQS. Also, as mentioned by Jason earlier in this thread, DQS enables you to cleanse your data before you can store and manage it as master data using MDS.

    I am sorry that you could not find the information that you were looking for. If you are loking for some specific information/use case/business scenarios, please list them here, and we will be more than happy to provide information/guidance around it.

    Thanks,
    Vivek
    SQL Server Documentation

    2012年12月3日 下午 02:11
  • Hi Kumar, I was interested to place rightly MDS and DQS in a general business scenario and my idea could be good.

    Thanks

    2012年12月3日 下午 04:15
  • Hi,

    I need to know if it exists any documentations about the matching between the new Data Quality Services and MDS. I think that DQS represents a specific task of the MDM process, isn't it?

    Thanks


    Thanks for asking this question - the replies posted here were helpful to me.

    Paras Doshi (Blog: ParasDoshi.com | Twitter: @Paras_Doshi )

    2012年12月30日 上午 01:23