We would like to store and retrieve XML/Non-XML documents using a generated unique key rather than the filename or external info, which are potentially unreliable. Is there a way to do this without storing the key in an indexed field in the document, that is, without manipulating the document?
Note: we use an XML doc with information about the business context as a kind of catalog entry that points to the real document. Under a given context, there may be one or more documents.
Are there performance issues that must be considered when choosing a mechanism for identifying documents?
One possible solution would be to generate a guid and use this as part of the document name when you store it from your application. A document name can be assigned to xml documents as well as non-xml documents. This way it does not mean you have to add extra unique identification information to the documents being stored.
Thank you for your reply and apologies for my answering so late :(.
We’ll certainly try that out. However, we had information from SAG that using docname was only reliable when docs were stored via HTTP PUT. We’re using Tamino 4.4.1. Docs come in via a web service, so we use the Java API to add them to the database.