Joined Searches

The equivalent Java API query in TIBCO Patterns for the example mentioned in Joins Example, is:

NetricsJoin join_def = new NetricsJoin(“Persons”,
NetricsJoin.JOIN_FULL_AND_PARTIALS,
                                   false,
                                   null,
                                   String [] { “Addresses”, “Phones” }
                       );
NetricsQuery q_first_name = NetricsQuery.Simple(
                                 “John”,
                                    new String [] { “Persons.first_name”},
                                 null
                            );
NetricsQuery q_last_name =  NetricsQuery.Simple(
                                 “Smith”,
                                 new String [] {“Persons.last_name”},
                                 null
                            );
NetricsQuery q_street =     NetricsQuery.Simple(
                                 “123 Main St.”,
                                 new String [] {“Addresses.street”},
                                 null
                            );
NetricsQuery q_city =       NetricsQuery.Simple(
                                 “Clarksburg”,
                                 new String [] {“Addresses.city”},
                                 null
                            );
NetricsQuery q_phone_num =  NetricsQuery.Simple(
                                 “609-884-1010”,
                                 new String [] {“Phones.number”},
                                 null
                             );
NetricsQuery query = NetricsQuery.And(null,
                                      new NetricsQuery [] {
                                                q_first_name,
                                                q_last_name,
                                                q_street,
                                                q_city,
                                                q_phone_num
                                      }
                     );
NetricsSearchCfg search_def = new NetricsSearchCfg(join_def);
search_def.setNetricsQuery(query);

There is no equivalent to the WHERE clause expression of the SQL statement. A joined search must always have a parent table. It must also have at least one child table, and the child table must be a child of the given parent. However, it is not required to list all the child tables of the parent. Any subset of the child tables of the parent can be included in the joined search.

In a joined fuzzy search, the relationship between records is firmly fixed, it is only the matching of values in the records that is fuzzy. The relationship between records is fixed when the records are created.

•

A fuzzy joined search is always equivalent to a standard fuzzy search on a single table that is a fully denormalized version of the joined tables.

A fully denormalized table has all the fields from each of the individual tables. For example, consider the three tables to have the following records:

Persons Table
id	first_name	last_name	dob	SSN
1	john	smith	6/23/1963	012345678
2	john	smythe	5/23/1960	021345678
3	jim	myth	8/14/1984	918273645

Addresses Table
addr_id	person_id	street	city	state
1	1	123 Main St	Clarksburg	NJ
2	1	32 Route 33	Robbinsville	NJ
1	2	213 Main St	Clarksburg	NJ
3	1	312 State Rd	Princeton	NJ
1	3	231 Ferry Rd	Ewing	NJ

Phones Table
phone_id	person_id	number	type
1	1	609-833-1001	home
2	1	732-981-1100	cell
1	3	609-442-0101	cell
1		609-833-0011	none

The fully denormalized table:

Denormalized Table
id	first_ name	last_name	addr_ id	street	phone_ id	number	type
1	john	smith	1	123 Main St	1	609-833- 1001	home
1	john	smith	1	123 Main St	2	732-981- 1100	cell
1	john	smith	2	32 Route 33	1	609-833- 1001	home
1	john	smith	2	32 Route 33	2	732-981- 1100	cell
1	john	smith	3	312 State Rd	1	609-833- 1001	home
1	john	smith	3	312 State Rd	2	732-981- 100	cell
2	john	smythe	1	213 Main St
					1	609-883- 011	none
3	jim	myth	1	231 Ferry Rd	1	609-442- 0101	cell

There is a separate record for each possible combination of parent and child records. As there were three child records for “John Smith” in the Addresses table and two in the Phones table, you can see 6 separate records for “John Smith” in the de-normalized table. There are two rows that are not fully populated. One is a parent record that has no child in the “Phones” table, and the other is an orphan record in the “Phones” table.

Subtopics

Did you find this helpful?

Yes No