Search Stored Procedure Search

A stored procedure search allows the server to locate entities using a stored procedure in a database.

The stored procedure must provide the server with a list of the identifiers that it should use to update the users selection.

There are three possible ways that the stored procedure can accomplish this:

• return the list as the result of the stored procedure
• return the list as an output parameter
• populate a table

If a table and key are provided in the configuration then the third method will be used, otherwise one of the first two methods will be used. The second method will be used for Oracle (since it doesn't support the first method) and the first method for all other database (unless a special option, resultAsParameter, is set to indicate that the second method should be used for non-Oracle databases).

In all cases the parameters the user enters will be passed as input parameters to the stored procedure, in the order they are listed in the configuration.
It is up to the stored procedure to determine if the user did not provide a value for a particular parameter, as the server will always attempt to set a value for every parameter defined in the configuration. If the user did not supply a value then null will be passed (note that the administrator can specify that a value is required for each parameter in the configuration to ensure that values are populated with a non-null value).

Returning results directly

If the database supports returning results directly then the list of id's to use for the selection can be returned in this manner.

Returning results indirectly

If the database does not support returning results from a stored procedure then the results can be provided via an output parameter (and the stored procedure returns nothing). This is the default for Oracle (unless a table and key are set).

Returning results via a table

If the stored procedure wishes to return results via a table then the table and key parameters must be set in the configuration, and they will specify the table name and the column which contain the id's used to update the selection.

It is the responsibility of the stored procedure to clear any rows belonging to the userid that it does not wish to return. That is, the server will not clear any old records from the table before calling the stored procedure.

In addition to this the stored procedure must also store a userid in the table to indicate which rows in the table correspond to which user.
The value that should be stored in that column will be passed to the stored procedure as the first parameter (unless a special option, useridAtEnd, is set to indicate it is the last parameter). The other input parameters for the search will follow the userid.
Also, note that the name of the userid column can be changed by setting the useridColumn spacial option.

Namespace

urn:com.cohga.server.search.database#1.0

Tags

procedure

Properties

Sub-tags

Content

None

Notes

An ACL can either be defined in-line or referenced indirectly, but only one should be used (the in-line version will take priority).

Table and key must both be included if one is

parameter

Properties

Sub-tags

Notes

• If the controltype is listbox then a datadefinition must be provided, this datadefinition will provide the values to be displayed in the listbox.
• If the listbox datadefinition only contains 1 column then that column will supply both the label and the value. If it contains two values then the first column will supply the label and the second the value.
• If a parameter contains another parameter, then you are defining a cascading parameter where setting the first sub-parameter will enable, and filter, the second parameter, and setting the second parameter will do the same for the third, etc.
• In a cascading parameter all sub-parameters must be of listbox type, and the datadefinition should be set in the outer parameter, not the sub parameters.
• In a cascading parameter only one level of nesting should be used.
• In a cascading parameter the datadefinition should supply the columns for all of the parameters, and valuecolumn properties should be set for each sub-parameter, and labelcolumn should be also set for all parameters if a different label is to be displayed to the user.

Examples

Stored procedure that returns result directly
Assume a database that supports returning a selection (in this case SQL Server)

CREATE PROCEDURE RoadByNameDirect @roadname varchar(20) AS
  SELECT DISTINCT id FROM roads WHERE name=@roadname

Stored procedure that returns result as a parameter
In this case Oracle requires the REF CURSOR type to be defined in a separate package (rather than being referenced directly in the parameters). Well this is what I have found, YMMV, either way the parameter must be a reference to a cursor.

CREATE PACKAGE cursor_types AS
  TYPE cursorType IS REF CURSOR;
END;

CREATE PROCEDURE RoadByNameParameter(roadname IN roads.name%TYPE DEFAULT NULL, results OUT cursor_types.cursorType) IS
BEGIN
  OPEN results FOR
    SELECT DISTINCT id
    FROM roads
    WHERE name = roadname;
END;

Stored procedure that returns results in a table

CREATE PROCEDURE RoadByNameTable(user IN road_table.userid%TYPE, roadname IN roads.name%TYPE DEFAULT NULL) IS
BEGIN
  DELETE FROM road_table
    WHERE road_table.userid = user;

  INSERT INTO road_table(userid, id)
    SELECT userid, id
    FROM roads
    WHERE name = roadname;
  COMMIT;
END;

<search:procedure id="road_direct">
	<entity>road</entity>
	<displayName>by Stored Procedure (Direct)</displayName>
	<description>Stored procedure returns result directly</description>
	<dataSource>ds.main</dataSource>
	<procedure>RoadByNameDirect</procedure>
	<parameter id="roadname">
		<parameter:label>Road Name</parameter:label>
	</parameter>
</search:procedure>

<search:procedure id="road_parameter">
	<entity>road</entity>
	<displayName>by Stored Procedure (Parameter)</displayName>
	<description>Stored procedure returns result as output parameter</description>
	<dataSource>ds.main</dataSource>
	<procedure>RoadByNameParameter</procedure>
	<parameter id="roadname">
		<label>Road Name</label>
	</parameter>
</search:procedure>

<search:procedure id="road_table">
	<entity>road</entity>
	<displayName>by Stored Procedure (Table)</displayName>
	<description>Stored procedure populates a table</description>
	<dataSource>ds.main</dataSource>
	<procedure>RoadByNameTable</procedure>
	<table>road_table</table>
	<key>id</key>
	<parameter id="roadname">
		<label>Road Name</label>
	</parameter>
</search:procedure>