© 2001 The dbXML Group L.L.C.
This documentation is a work in progress. Links to the most current version can be found at http://www.dbxml.org/docs/
$Id: AdministratorsGuide.xml,v 1.13 2001/09/12 18:47:32 kstaken Exp $
Database administration of dbXML is accomplished from the command line using the dbxmladmin command. This command allows you to view and alter the database configuration on the fly on a running system. A complete list of available commands and more detail about each command can be found in the Command Line Tools Reference Guide.
Adds a collection named products under the collection /db/data.
dbxmladmin add_collection -c /db/data -n products
Deletes the collection named products from the collection /db/data.
dbxmladmin delete_collection -c /db/data/products
The dbXML indexing system allows you to define indexes to speed performance on commonly used XPath queries. If no indexes are defined you can still execute queries but performance will suffer because the query engine will need to scan the entire collection to create the result node-set.
Indexes can be added using the dbxmladmin command.
Using this simple XML file you might want to index the product_id element because searches for products by product_id are common.
<?xml version="1.0"?> <product> <product_id>120320</product_id> <description>Glazed Ham</description> </product>
This can be accomplished by running the following command. This will create an index named idindex on all product_id elements in the collection /db/data/catalog.
dbxmladmin add_indexer -c /db/data/catalog -n idindex -p product_id
Once this is done the query engine will now use this index to help resolve XPath queries that involve restriction on the value of the product_id element.
The -p parameter to the command specifies the pattern to use in the index. These patterns are used by the Indexing system to determine best-fit and match-based Indexers for queries and index updating. The pattern used MUST resemble the following scheme.
Pattern Description =========== ==================================================== elem The value of the named element elem@attr The value of the attribute for the named element * The value for all elements *@attr The value of the named attribute for all elements elem@* The value of all attributes for the named element *@* The value of all attributes for all elements
Note: In order to index a namespace other than the default namespace, you must prepend your pattern components with a URI placed in square brackets. Example:
[http://www.world.org/People]person *@[http://www.world.org/People]id [http://www.world.org/People]person@[http://www.world.org/People]id
Do not include a prefix in these patterns, as the indexing system, like most Namespace processing applications, processes namespaced elements and attributes independently of the prefix that is used.
1.5. Indexing both Elements and Attributes
Because the patterns recognize either an element or an attribute, and not both, in order to index all element and attribute values in a collection, you'd have to create two index entries. The * pattern will index all elements and the *@* pattern will index all attributes of all elements.
dbxmladmin add_collection_indexer -c /db/data/catalog -n idindex -p '*' dbxmladmin add_collection_indexer -c /db/data/catalog -n idindex -p '*@*'
Excessive use of wildcard indexes can adversely affect the performance of the indexing system. Best practice would be to use specific element or attribute indexes whenever possible, and only define wildcard indexes when it is absolutely necessary.
The server must be started from within the dbXML-Core directory. A future revision of the server will fix this limitation.
The dbXML server can be easily shutdown from the command line. You must provide the name of the server instance to shutdown.
Currently backing up dbXML consists of simply shutting down the server and copying the entire contents of the dbXML-Core/db directory to the backup media. [1]
This example assumes that the dbXML-Core/bin directory is in your path.
cd dbXML-Core dbxmladmin shutdown cp -pr db /backup/db ./start
[1] A more robust online backup operation will need to be developed in the future