Daniel Bloomfield Ramagem wrote:
I'm a little confused about indexing in Jackrabbit. Indexing is only
useful
if I want to be able to perform searches of the *contents* of the nodes in
the repository, right? Will not having an index still allow XPath queries?
indexing covers both fulltext indexing as well as indexing of basic property
types like string, long, date, etc. what kind of resources are fulltext indexed
is configurable.
I have a Jackrabbit repository set up that I want to use as a virtual file
system. So I have a bunch of nt:file and nt:folder node types. I don't
care about searching inside the contents of the files stored in the nt:file
node (via "jcr:data").
According to http://jackrabbit.apache.org/doc/config.html the <SearchIndex>
element of the repository.xml is optional and if omitted the workspace is
not searchable. I've commented out this section in my configuration but
then I get a RepositoryException "no search manager configured for this
workspace".
this means you disabled the search manager entirely. a call to
Workspace.getQueryManager() will then throw the exception you mentioned.
If you are not interested in fulltext indexing nt:resource nodes you should
actually be fine with the default configuration. only if you want to index
certain document types like pdf you need to add configuration parameters to the
searchindex tag. See:
http://svn.apache.org/repos/asf/jackrabbit/tags/1.2.1/jackrabbit-index-filters/README.txt
regards
marcel
