original
In the end, the hand-output XML object is reliable from January 06, 2005 and XML, it is often crying: Why do I spend so much vigorous and over the XML document? In fact, I just want to access one of the objects of one of the objects! ! I learned that Castor felt this is a solution, but it still needs to read and write updates throughout the document. One is the time limit does not allow the current in-depth study, and that is also a project that is not a mature project; the second, the adoption will conflict with the current Digester read mode, it is a bit. However, the update frequency set by the pictures of the pictures of the pictures seems to be higher and higher, and then read only by hand in XML. It doesn't work. Therefore, it is intended to transplant the department object into the database, but once a job, I find that it is not cost-effective.
In fact, it is true that Ignore a problem, in fact, a contrast of XML and relational data inventory storage: The means of access to the XML is indeed very mature, but it is stored in the hierarchy of the object, but the relational database is Flat formally stored. I am currently going to use Secion to convert to a relational database, the purpose is to be more convenient to make the chamber change settings, which is a file in XML, but it is a relational database, but a bunch of relationship tables (one is also a relationship concept) Entity, but this entity is non-inspector, which means a bunch of response to an object), and is associated with the like form, etc., quite complex. This may not be a good idea, and in large projects, the use of complex relationships is not much recorded, which seems to be a cost-effective approach to a low-time approach. So I hesitated.
Another way is to do an updateable XML module: handling techniques include:
1. Modify SectionBase so that it is a number of XML work for multiple departments, not like everything now, reading all the departments;
2, make a way to update the XML of each department; whether it is CASTO, or other methods;
3, make an interface to update each department, connect it to the department management station.
The key here is the second step. In order to confirm that the second step can be done in the current simplest method, then the previously downloaded article, but the blog China is getting weaker, no ban, actually, I can't stop it, I have a long time to open the original article. Read one again. The results of the study are still the same. If CaStor means uses its JDO, not just XML output, and my reading is mainly used by Digester; so here contains a wider range of modifications, and contains A more uncertain trial; this is why I have been temporarily released in two weeks: there is no time to study its use ideas and practical applications. It seems that only SAX or even string of the original point can be used.
Consider the common SAX / Xalan / Jaxp / JDOM several processing techniques, if not working with non-objective document content, you need to write an XSLT / and converter, regardless of how, to have a consistent Java object / XML object matches the mode work, but also to make the above document objects can be interchanged with the Java object after Digester: Have you heard of it! ! Without a few people from Digester, I really have used it, I'm thinking that the forum asked the old brothers outside the country. I think this way, if there is, I must get an object from the JDOM in Digester, and re-turned to Document objects - but no! !
It seems that I have to consider how the self-achievable and reusability is. ... I want to come down, in fact, this is not a difficult thing, just need to implement an interface in each class, such as Write, then call the step-by-step call? Why do you want to be far away, find some unreliable things trial? One through Baitong, actually manually output the object string is not a horror thing, I am letting the authors of the article, the key is that this is output according to the object, the amount is not big, and it is also quite Good management. Little means compared with the servlet output. The original method is not necessarily behind, it is appropriate!
