Metacat

/**

 *  '$RCSfile$'

 *  Copyright: 2004 Regents of the University of California and the

 *             National Center for Ecological Analysis and Synthesis

 *

 *   '$Author$'

 *     '$Date$'

 * '$Revision$'

 *

 *

 * This program is distributed in the hope that it will be useful,

 * GNU General Public License for more details.

 *

 * You should have received a copy of the GNU General Public License

 */

 * xmltables.sql except that this script is to be use to

 * create the database tables on a Postgresql backend rather

 * than an Oracle Backend

/*

CREATE SEQUENCE xml_replication_id_seq;

CREATE TABLE xml_replication (

  last_checked DATE,

  replicate INT8,

  datareplicate INT8,

  hub INT8,

  CONSTRAINT xml_replication_pk PRIMARY KEY (serverid)

 */

CREATE SEQUENCE xml_nodes_id_seq;

CREATE TABLE xml_nodes (

	nodeid INT8 default nextval('xml_nodes_id_seq'),

					-- the unique node id (pk)

	nodeindex INT8,		-- order of nodes within parent

	nodetype VARCHAR(20),	-- type (DOCUMENT, COMMENT, PI,

        nodeprefix VARCHAR(50), -- the namespace prefix of the node

	parentnodeid INT8,	-- index of the parent of this node

	rootnodeid INT8,	-- index of the root node of this tree

	date_created DATE,

	date_updated DATE,

    nodedatadate TIMESTAMP, -- the data for this node if it is a date

);

 */

CREATE INDEX xml_nodes_idx1 ON xml_nodes (rootnodeid);

CREATE INDEX xml_nodes_idx2 ON xml_nodes (parentnodeid);

CREATE INDEX xml_nodes_idx3 ON xml_nodes (nodename);

 */

CREATE TABLE xml_nodes_revisions (

        nodeid INT8,            -- the unique node id (pk)

        nodeindex INT8,         -- order of nodes within parent

        nodetype VARCHAR(20),   -- type (DOCUMENT, COMMENT, PI,

                                -- ELEMENT, ATTRIBUTE, TEXT)

        nodename VARCHAR(250),  -- the name of an element or attribute

        nodeprefix VARCHAR(50), -- the namespace prefix of the node

        parentnodeid INT8,      -- index of the parent of this node

        rootnodeid INT8,        -- index of the root node of this tree

        docid VARCHAR(250),     -- index to the document id

        date_created DATE,

        date_updated DATE,

        nodedatadate TIMESTAMP, -- the data for this node if it is a date

   CONSTRAINT xml_nodes_revisions_root_fk

                FOREIGN KEY (rootnodeid) REFERENCES xml_nodes_revisions,

   CONSTRAINT xml_nodes_revisions_parent_fk

                FOREIGN KEY (parentnodeid) REFERENCES xml_nodes_revisions

);

/*

 * Indexes of rootnodeid, parentnodeid, and nodename in xml_nodes_revisions

 */

CREATE INDEX xml_nodes_revisions_idx1 ON xml_nodes_revisions (rootnodeid);

CREATE INDEX xml_nodes_revisions_idx2 ON xml_nodes_revisions (parentnodeid);

CREATE INDEX xml_nodes_revisions_idx3 ON xml_nodes_revisions (nodename);

/*

 */

CREATE SEQUENCE xml_catalog_id_seq;

CREATE TABLE xml_catalog (

	catalog_id INT8 default nextval('xml_catalog_id_seq'),

                                        -- the id for this catalog entry

	entry_type VARCHAR(500),	-- the type of this catalog entry

					-- (e.g., DTD, XSD, XSL)

	source_doctype VARCHAR(500),	-- the source public_id for transforms

	target_doctype VARCHAR(500),	-- the target public_id for transforms

	public_id VARCHAR(500),	-- the unique id for this type

	system_id VARCHAR(1000),	-- the local location of the object

 */

CREATE SEQUENCE xml_documents_id_seq;

 */

CREATE TABLE xml_documents (

	docid VARCHAR(250),	-- the document id #

	rootnodeid INT8,		-- reference to root node of the DOM

	docname VARCHAR(100),	-- usually the root element name

	doctype VARCHAR(100),	-- public id indicating document type

	user_owner VARCHAR(100),	-- the user owned the document

	user_updated VARCHAR(100),	-- the user updated the document

	server_location INT8,	-- the server on which this document resides

	rev INT8 default 1,   -- the revision number of the document

	date_created DATE,

	date_updated DATE,

	public_access INT8,	-- flag for public access

     CONSTRAINT xml_documents_rep_fk

    CONSTRAINT xml_documents_root_fk

);

 */

CREATE INDEX xml_documents_idx1 ON xml_documents (docid, doctype);

 *                    or DELETE

 */

CREATE SEQUENCE xml_revisions_id_seq;

CREATE TABLE xml_revisions (

	revisionid INT8  default nextval('xml_revisions_id_seq'),

                                        -- the revision number we are saving

	docid VARCHAR(250),	-- the document id #

	rootnodeid INT8,		-- reference to root node of the DOM

	docname VARCHAR(100),	-- usually the root element name

	doctype VARCHAR(100),	-- public id indicating document type

	user_owner VARCHAR(100),

	user_updated VARCHAR(100),

	server_location INT8,

	rev INT8,

	date_created DATE,

	date_updated DATE,

	public_access INT8,	-- flag for public access

   CONSTRAINT xml_revisions_rep_fk

		FOREIGN KEY (server_location) REFERENCES xml_replication,

);

 */

CREATE TABLE xml_access (

	permission INT8,		-- "read", "write", "all"

	perm_type VARCHAR(32),	-- "allowed" or "denied"

	perm_order VARCHAR(32),	-- "allow first" or "deny first"

	begin_time DATE,		-- the time that permission begins

	end_time DATE,		-- the time that permission ends

	ticket_count INT8,		-- ticket counter for that permission

  subtreeid VARCHAR(32),

  startnodeid INT8,

  endnodeid INT8,

CREATE INDEX xml_access_idx2 ON xml_access (permission);

CREATE INDEX xml_access_idx3 ON xml_access (perm_type);

CREATE INDEX xml_access_idx4 ON xml_access (perm_order);

CREATE INDEX xml_access_idx5 ON xml_access (subtreeid);

 * ALTER TABLE xml_access ADD COLUMN guid text;

*/

 * Index of Nodes -- table to store precomputed paths through tree for

 */

CREATE TABLE xml_index (

	nodeid INT8,		-- the unique node id

	doctype VARCHAR(100),	-- public id indicating document type

				-- by this row

   CONSTRAINT xml_index_nodeid_fk FOREIGN KEY (nodeid) REFERENCES xml_nodes,

);

 * Index of the paths in xml_index

CREATE INDEX xml_index_idx1 ON xml_index (path);

CREATE SEQUENCE xml_relation_id_seq;

CREATE TABLE xml_relation (

	relationid INT8 default nextval('xml_relation_id_seq') PRIMARY KEY,

					     -- unique id

	docid VARCHAR(250) ,         -- the docid of the package file

	                                     -- that this relation came from

        packagetype VARCHAR(250),          -- the type of the package

	subject VARCHAR(512) NOT NULL, -- the subject of the relation

	subdoctype VARCHAR(128),         	-- the doctype of the subject

	relationship VARCHAR(128)  NOT NULL,-- the relationship type

	object VARCHAR(512) NOT NULL, -- the object of the relation

	objdoctype VARCHAR(128),          -- the doctype of the object

	CONSTRAINT xml_relation_uk UNIQUE (docid, subject, relationship, object),

);

 * GUID, and a local metacat identifier, which consists of the docid

 * and revision fields. Each row maps one global identifier to the local

 * identifier (docid) used within metacat.

   docid  VARCHAR(250),	 -- the local document id #

   rev    INT8,          -- the revision part of the local identifier

   CONSTRAINT identifier_pk PRIMARY KEY (guid)

CREATE INDEX identifier_docid on identifier(docid);

CREATE INDEX identifier_rev on identifier(rev);

CREATE INDEX identifier_docid_rev on identifier(docid, rev);

 * similar restraints to identifier.  Cannot use identifier table for this

 * purpose because then you have to worry about whether you insert the

 * data first or the systemMetadata first.

 */

CREATE TABLE systemMetadata (

	rights_holder VARCHAR(250), --the user who has rights to the document, usually the first persons to upload it

	checksum VARCHAR(512), --the checksum of the doc using the given algorithm (see below)

	checksum_algorithm VARCHAR(250), --the algorithm used to calculate the checksum

	origin_member_node VARCHAR(250), --the member node where the document was first uploaded

	authoritive_member_node VARCHAR(250), --the member node that currently controls the document

	date_modified TIMESTAMP, -- the last date/time that the file was changed

	submitter VARCHAR(256), -- the user who originally submitted the doc

	object_format VARCHAR(256), --the format of the object

	size VARCHAR(256), --the size of the object

	number_replicas INT8, 	-- the number of replicas allowed

	obsoleted_by   text,     -- the identifier of the record that replaces this record

  file_name    text,      -- the suggested file name for this object

 * can have multiple properties, we have to store them in a separate table. The guids in this table refer

CREATE TABLE smMediaTypeProperties (

	guid    text,  -- id refer to guid in the system metadata table

  name    text, -- name of the property

  value    text, -- value of the property

);

/*

 * ALTER TABLE systemMetadata DROP COLUMN docid;

 * ALTER TABLE systemMetadata DROP COLUMN rev;

 * ALTER TABLE systemMetadata ADD COLUMN number_replicas INT8;

	member_node VARCHAR(250),	 -- replication member node

	policy text,	 -- the policy (preferred, blocked, etc...TBD)

	member_node VARCHAR(250),	 -- replication member node

	status VARCHAR(250),	 -- replication status

	date_verified TIMESTAMP, 	-- the date replication was verified

CREATE TABLE xml_accesssubtree (

	docid		VARCHAR(250),	-- the document id #

  rev 		INT8 default 1, --the revision number of the docume

  controllevel VARCHAR(50), -- the level it control -- document or subtree

  endnodeid INT8, -- the end node if of access subtree

);

/*

 *		    and the number of times this table is accessed

 */

        returnfield_id     INT8 default nextval('xml_returnfield_id_seq'),   -- the id for this returnfield entry

        returnfield_string VARCHAR(2000),                                    -- the returnfield string

        usage_count        INT8,                                             -- the number of times this string has been requested

        CONSTRAINT xml_returnfield_pk PRIMARY KEY (returnfield_id)

);

CREATE INDEX xml_returnfield_idx1 ON xml_returnfield(returnfield_string);

/*

 * Queryresults -- table to store queryresults for a given docid

 * and returnfield_id

 */

  queryresult_id INT8 default nextval('xml_queryresult_id_seq'), -- id for this entry

  returnfield_id       INT8,          -- id for the returnfield corresponding to this entry

  docid                VARCHAR(250),  -- docid of the document

  CONSTRAINT xml_queryresult_pk PRIMARY KEY (queryresult_id),

  CONSTRAINT xml_queryresult_searchid_fk

               FOREIGN KEY (returnfield_id) REFERENCES xml_returnfield

);

CREATE INDEX xml_queryresult_idx1 ON xml_queryresult (returnfield_id, docid);

/*

 */

CREATE SEQUENCE access_log_id_seq;

CREATE TABLE access_log (

  entryid       INT8 default nextval ('access_log_id_seq'), -- the identifier for the log event

  ip_address    VARCHAR(512),   -- the ip address inititiating the event

  principal     VARCHAR(512),   -- the user initiating the event

  event         VARCHAR(512),   -- the code symbolizing the event type

  date_logged   TIMESTAMP,      -- the datetime on which the event occurred

  CONSTRAINT access_log_pk PRIMARY KEY (entryid)

);

 * the index_event table for solr-based indexing

 */

CREATE TABLE index_event (

	guid text,

	event_action VARCHAR(250),

	description text,

	event_date TIMESTAMP

);

 */

CREATE SEQUENCE xml_path_index_id_seq;

CREATE TABLE xml_path_index (

    nodeid INT8  default nextval('xml_path_index_id_seq'),

        docid VARCHAR(250),     -- the document id

        nodedatadate TIMESTAMP, -- the data for this node if it is a date

                                -- by this row

   CONSTRAINT xml_path_index_pk PRIMARY KEY (nodeid),

);

/*

 * Indexes of path, nodedata and nodedatanumerical in xml_path_index

 */

CREATE INDEX xml_path_index_idx1 ON xml_path_index (path);

CREATE INDEX xml_path_index_idx2 ON xml_path_index (nodedata);

CREATE INDEX xml_path_index_idx3 ON xml_path_index (nodedatanumerical);

/*

 */

CREATE TABLE harvest_site_schedule (

  site_schedule_id INT8,         -- unique id

  documentlisturl  VARCHAR(255), -- URL of the site harvest document list

  ldapdn           VARCHAR(255), -- LDAP distinguished name for site account

  datenextharvest  DATE,         -- scheduled date of next harvest

  datelastharvest  DATE,         -- recorded date of last harvest

  updatefrequency  INT8,         -- the harvest update frequency

  unit             VARCHAR(50),  -- update unit -- days weeks or months

  contact_email    VARCHAR(50),  -- email address of the site contact person

  ldappwd          VARCHAR(20),  -- LDAP password for site account

  CONSTRAINT harvest_site_schedule_pk PRIMARY KEY (site_schedule_id)

);

 */

CREATE TABLE harvest_log (

  harvest_log_id         INT8,          -- unique id

  harvest_date           DATE,          -- date of the current harvest

  status                 INT8,          -- non-zero indicates an error status

  message                VARCHAR(1000), -- text message for this log entry

  CONSTRAINT harvest_log_pk PRIMARY KEY (harvest_log_id)

);

 *                       generated errors during the harvest

 */

CREATE TABLE harvest_detail_log (

  detail_log_id          INT8,          -- unique id

  harvest_log_id         INT8,          -- ponter to the related log entry

  scope                  VARCHAR(50),   -- document scope

  identifier             INT8,          -- document identifier

  revision               INT8,          -- document revision

  document_url           VARCHAR(255),  -- document URL

  error_message          VARCHAR(1000), -- text error message

  document_type          VARCHAR(100),  -- document type

  CONSTRAINT harvest_detail_log_pk PRIMARY KEY (detail_log_id),

);

 * db_version -- table to store the version history of this database

 */

CREATE SEQUENCE db_version_id_seq;

CREATE TABLE db_version (

  db_version_id   INT8 default nextval ('db_version_id_seq'), -- the identifier for the version

  version         VARCHAR(250),     -- the version number

  status          INT8,             -- status of the version

  date_created    TIMESTAMP,        -- the datetime on which the version was created

  CONSTRAINT db_version_pk PRIMARY KEY (db_version_id)

);

/*

 * scheduled_job -- table to store scheduled jobs

 */

CREATE SEQUENCE scheduled_job_id_seq;

CREATE TABLE scheduled_job (

  id INT8 NOT NULL default nextval('scheduled_job_id_seq'),

  date_created TIMESTAMP NOT NULL,

  date_updated TIMESTAMP NOT NULL,

  status VARCHAR(64) NOT NULL,

  name VARCHAR(512) NOT NULL,

  trigger_name VARCHAR(512) NOT NULL,

  group_name VARCHAR(512) NOT NULL,

  class_name VARCHAR(1024) NOT NULL,

  start_time TIMESTAMP NOT NULL,

  interval_unit VARCHAR(8) NOT NULL,

  CONSTRAINT scheduled_job_pk PRIMARY KEY (id),

  CONSTRAINT scheduled_job_uk UNIQUE (name)

);

/*

 * scheduled_job_params -- table to store scheduled jobs

 */

CREATE SEQUENCE scheduled_job_params_id_seq;

CREATE TABLE scheduled_job_params (

  id INT8  NOT NULL default nextval('scheduled_job_params_id_seq'),

  date_created TIMESTAMP NOT NULL,

  date_updated TIMESTAMP  NOT NULL,

  status VARCHAR(64)  NOT NULL,

  job_id INT8 NOT NULL,

  key VARCHAR(64) NOT NULL,

  value VARCHAR(1024) NOT NULL,

  CONSTRAINT scheduled_job_params_pk PRIMARY KEY (id),

  CONSTRAINT scheduled_job_params_fk

        FOREIGN KEY (job_id) REFERENCES scheduled_job(id)

);