Hi Mansi, my 2 cents here are mainly that none of the RDBMS based solution deliver the full support of the complete stack of XML technologies to their complete power (XPath, XQuery 3.x, XSLT, XSD, eXPath , aso). As far as I remember you can hardly find HTTP server functionality with RestXQ support there too. As soon as you have to write an application or service of a respectable complexity you will nee all of those functionalities. Another point to use here, at least with BaseX, is the footprint of the application. I don't know how large DB2 is but for sure installing, managing and using PostgreSQL is several levels of complexity higher than using BaseX. Finally I also think that for implementing a really performing tree based database the relational technology is not suited and a proper datamodel is required but I don't have figures or demonstrations for that so I'm also very curious to see those performance comparison results! :-D Hope this helps somehow. Greetings, Marco.
On 28/02/2015 12:11, Christian Grün wrote:
Hi Mansi,
No one answered so far, so there is probably no simple answer to your question.
As you indicated, many relational DBMS (such as DB2 and PostgreSQL) provide support for XPath as well, but in most cases, the systems do not create additional index structures on top of the XML nodes to speed up querying. However, as the queries you presented on the mailing list are simple XPath expressions, there are probably not so many chances to benefit from values index structures anyway.
If teams in your companies are trying DB2 and PostgreSQL anyway, I'll be looking forward to hearing more about their experiences. Commercial systems often disallow their users to publish performance results (and presenting them on the list is some way of publishing them), so you can also send them to me privately. If you go for PostgreSQL, this won't be a problem of course.
Hope this helps, Christian