canasia.org

Hello.   I would like to build an application which is used to store a lot of hierarchical data, which has a constantly changing data structure, and in which the same piece of data might be simultaneously accessed or modified by more than one user.  Obviously, _xml_ is well-suited to representing hierarchical data, and _xml_ schemas are well suited to dynamic data structures.  However, the _xml_ tools provided with D6E allow for basically a flatfile data_base_, and it isn't clear to me whether or not it is a true data_base_ management system that could handle large quantities of data and multi-threaded reads/writes. So what I am trying to decide is whether or not to use Delph6Ent with just its existing _xml_ tools, to employ some kind of third party _xml_ data_base_, or to use _xml_ in conjunction with an RDBMS like Inter_base_ (Firebird), in which case I would use the RDBMS to keep track of all of the _xml_ files.  Any suggestions on this? As an alternative (perhaps this is considered heresy on this newsgroup), I am considering the use of Python along with an _xml_ data_base_ such as either eXact or Xindice (both open source).  Does anyone know if eXact or Xindice can be used with D6E?  Is there a Delphi API for either of those _xml_ data_base_ systems? I would greatly appreciate any insights or recommendations anyone could provide?  I am also interested in any experiences related to what I am trying to do. Thank you, Peter

When you endorse DOM-_base_d access to the heirarchical data, are you suggesting something different than the Delphi _xml_ components that ship with D6E?  If so, what do you recommend? err, I'm not sure. I presumed you meant _xml_ _base_d client dataset stuff with the mapper. To me that didn't meet your goals which seem quite similar to mine - a set of quite discrete records that are highly (and flexibly) structured, along with a set of advanced (and flexible) search requirements that fall outside what can easily be achieved using SQL (for instance, I need to do search by NYSIIS instead of soundex and also use Jaro-Winkler for improving search - though I'm not sure what that means) I'm using a combination of ms_xml_ and open_xml_ and I will also use the SAX package, along with my standard sql _base_d data access (ODBCExpress, since it never fails) There's lot's of ways to do this - maybe the D6 components offer some help, but they seem to be very VB like to me - good if you want simple, but not good if you want to do it properly. I'm also evaluating tiopf for my data _layer_ Grahame

the _xml_ data_base_s certainly provide some of that functionality and are really useful where you want the data_base_ to store the _xml_ in some structured format to allow you to slice and dice it, and search across records using XPath Agreed. Another good point to make is internal optimization. Your data should definitely influence you decision. If you are looking at large (e.g. 100K+) chunks of _xml_ then _xml_DBS may not be for you. Most are designed for small fragments of _xml_ <= 10K

There's lot's of ways to do this - maybe the D6 components offer some help, but they seem to be very VB like to me - good if you want simple, but not good if you want to do it properly. I'm also evaluating tiopf for my data _layer_ Grahame toipf looks interesting, but I'm not sure I have the time to delve into such an approach. Thanks for your comments. -P