Fabian, Jens,
thanks you very much for providing Swati (and all other readers of this list) with the detailed information.
Swati (and others),
here's the "missing" bit:
o Upon each access to a cached document, MonetDB will try to figure out if the document has changed since caching it. In particular, it verifies file size and timestamp for documents in the local filesystem. (I don't know, actually, what the caching system does for documents from the network.)
======== # Monet Database Server V4.10.3 # Copyright (c) 1993-2006, CWI. All rights reserved. # Compiled for x86_64-redhat-linux-gnu/64bit with 64bit OIDs; dynamically linked. # Visit http://monetdb.cwi.nl/ for further information. MonetDB>module(pathfinder); MonetDB>sigs("pathfinder"); #---------------------------------------------------------# # signature # name # str # type #---------------------------------------------------------# [ "delete_all_docs(bit) : void" ] [ "delete_doc(str) : void" ] [ "pfstart(bit, lng) : void" ] [ "shred_doc(str, str) : void" ] [ "xmlcache_add_rule(str, any) : void" ] [ "xmlcache_del_rule(str) : void" ] [ "xmlcache_print() : void" ] [ "xmlcache_print_rules() : void" ] [ "xmldb_print() : void" ] [ "xquery(mode str, xquery str, is_url bit) : str " ] [ "xquery_server(Stream in, Stream out) : void " ] [ "xquery_start_query_cache(lng) : ptr" ] MonetDB>help("xmlcache_add_rule"); PROC: xmlcache_add_rule(str, lng) : void no text available
PROC: xmlcache_add_rule(str, any) : void MODULE: pathfinder COMPILED: by boncz on May 2005 DESCRIPTION: add a new URI lifetime rule.
The XML document cache keeps indexed copies of documents that where recently used in the fn:doc(URI) xquery function.
The size of the cache is controlled using the 'xquery_cacheMB' setting in the 'MonetDB.conf' file.
For file URIs, the cache looks at the last-modification-time of the file on disk to guarantee that the cached document is still up-to-date for answering queries from.
For other URIs, *lifetime rules* determine how long documents can stay in the cache. Each lifetime rule consists of a URI prefix and the registered seconds of lifetime.
The rule with longest prefix that matches an URI counts. Specifying a lifetime of 'int(nil)' seconds means that the URI will *not* be cached at all. This is also the default if no prefix matches an URI.
The name of a cached document is the same as its location (URI). For explicitly shredded documents (with 'shred_doc(location,name)'), the name is an 'alias' and may differ from the URI. Explicitly shredded documents fall outside the XML document cache; documents are only removed at explicit user request (with 'delete_doc(name)').
MonetDB> ========
Cheers,
Stefan