Open opeeters opened 12 years ago
I will support and try to prioritise this feature request. It would lower the entrance barrier for advanced dynamic data use cases for daviz, as most structured data is kept in closed relational databases.
Some info why we did not include this in the start: With daviz we have focused and promoted a "web of data" architecture, where data visualization tools were totally lacking. (For relational databases the market already offers many solutions e.g. the open source "pentaho") The goal of eea.daviz is not only to help visualize data TTW, but also to open the users to the large amount of data already available in the linked data clouds, and hopefully enlight them about the benefits of opening up their databases with data web standards set by W3C.
As discussed in the ploneconf there still too many of our clients who are not familiar with RDF and still haven´t seen its advantages since they are working only closed internal databases or using dedicated web services APIs (web 2.0). However, APIs are good and solving many use cases, they have many drawbacks and issues that Linked Data would have solved, see a recent article from programmableweb about this "Linked Data to Take Programmable Web to a Higher Level" http://blog.programmableweb.com/2012/07/24/linked-data-to-take-programmable-web-to-a-higher-level/
Adding ODBC connectivity to eea.daviz would help users visualise their closed data with zope/plone, but will not solve the data-silo syndrome. We hope though that playing with eea.daviz would make these new users more aware of the open linked data on the web.
This is the plone product we could use as part of the solution http://plone.org/products/mxodbc-zope-da
I fully support the development of a semantic web. Daviz is a super tool to support this development. From the perspective of Web 2.0 I don't however see any difference between supporting csv/tsv or copy/paste and ODBC connectivity. The latter will certainly help an additional potential user group to look beyond their silo.
One question on performance though... Transmitted over the WWW, JSON versus XML... JSON wins for sure. But if you want to visualise dynamic data (eg complex latest hourly values) a direct DB-connection might still be more efficient than even a SPARQL endpoint, not?
@alecghica @demarant @avoinea It would be great addition to this wonderful tool if it would also be possible to connect to ODBC databases and pull data out of any compliant DB-architecture via a SQL-statement. If you can look at how you could pull this data into daviz, I can help with the documentation and a generic database connection.