Query System in Aletyx Enterprise Build of Drools

Introduction to Drools Queries

The Query system in Aletyx Enterprise Build of Drools provides a powerful mechanism for extracting and analyzing data from the rule engine. Queries allow developers to retrieve facts from working memory that match specific conditions, without executing actions on those facts. In the context of Rule Units, queries become even more powerful as they can be scoped to specific rule units. Queries function as a bridge between the rule engine and the application, enabling a more interactive approach to working with the knowledge base.

Traditional Queries vs. Rule Unit Queries

Aletyx Enterprise Build of Drools supports two types of query approaches:

Traditional Queries

Traditional queries operate on the global working memory and are not attached to any specific rule unit:

package org.example;

import org.example.Customer;
import org.example.Account;

query "FindPremiumCustomers"
    $customer: Customer(status == "premium")
    $account: Account(customerId == $customer.id, balance > 10000)
end

Rule Unit Queries

Rule Unit queries are scoped to a specific Rule Unit and operate only on the data sources within that unit:

package org.example;

unit CustomerRuleUnit;

import org.example.Customer;
import org.example.Account;

query "FindPremiumCustomers"
    $customer: /customers[status == "premium"]
    $account: /accounts[customerId == $customer.id, balance > 10000]
end

Note the key differences:

The unit declaration at the top
The OOPath-like syntax (/customers[...]) for accessing data sources

Query Definition

Basic Structure

A query definition consists of:

A unique name
Optional parameters
A set of patterns to match against facts
For Rule Unit queries, reference to the specific Rule Unit

query "QueryName" (Type1 param1, Type2 param2, ...)
    // Patterns to match facts
end

Pattern Matching in Queries

Queries use the same pattern matching syntax as rules, allowing for complex conditions and constraints:

query "ComplexPatternQuery"
    // Simple pattern
    $customer: Customer(age > 21)

    // Pattern with multiple constraints
    $account: Account(
        customerId == $customer.id,
        status != "closed",
        balance > 1000
    )

    // Pattern with nested property access
    $order: Order(
        customer.id == $customer.id,
        items.size > 0,
        total > 100
    )
end

Parameters in Queries

Queries can accept parameters, making them more flexible and reusable:

query "FindCustomersByAgeRange" (int minAge, int maxAge)
    $customer: Customer(age >= minAge, age <= maxAge)
end

In Rule Units, these parameters are used similarly:

query "FindCustomersByAgeRange" (int minAge, int maxAge)
    $customer: /customers[age >= minAge, age <= maxAge]
end

Executing Queries

In Traditional Drools Applications

KieSession kieSession = kieContainer.newKieSession();
// Insert facts
kieSession.insert(new Customer("C1", "John", 30, "regular"));
kieSession.insert(new Customer("C2", "Mary", 25, "premium"));

// Execute query without parameters
QueryResults results = kieSession.getQueryResults("FindPremiumCustomers");

// Execute query with parameters
QueryResults ageResults = kieSession.getQueryResults("FindCustomersByAgeRange", 18, 25);

// Process results
for (QueryResultsRow row : results) {
    Customer customer = (Customer) row.get("$customer");
    Account account = (Account) row.get("$account");
    System.out.println(customer.getName() + " has balance: " + account.getBalance());
}

In Rule Unit Applications

// Create and populate Rule Unit
CustomerRuleUnit customerUnit = new CustomerRuleUnit();
customerUnit.getCustomers().insert(new Customer("C1", "John", 30, "regular"));
customerUnit.getCustomers().insert(new Customer("C2", "Mary", 25, "premium"));
customerUnit.getAccounts().insert(new Account("A1", "C1", 20000));
customerUnit.getAccounts().insert(new Account("A2", "C2", 15000));

// Create Rule Unit Executor
RuleUnitExecutor executor = RuleUnitExecutor.create().bind(kieContainer);

// Execute query without parameters
QueryResults results = executor.run(customerUnit, "FindPremiumCustomers");

// Execute query with parameters
QueryResults ageResults = executor.run(customerUnit, "FindCustomersByAgeRange", 18, 25);

// Process results (same as traditional)
for (QueryResultsRow row : results) {
    Customer customer = (Customer) row.get("$customer");
    Account account = (Account) row.get("$account");
    System.out.println(customer.getName() + " has balance: " + account.getBalance());
}

Advanced Query Features

Nested Queries

Queries can reference other queries, allowing for composition and reuse:

query "ActiveCustomers"
    $customer: /customers[status == "active"]
end

query "PremiumActiveCustomers"
    ActiveCustomers($customer;)
    /accounts[customerId == $customer.id, balance > 10000]
end

The syntax ActiveCustomers($customer;) passes the $customer variable from the nested query to the outer query.

Backward Chaining

Aletyx Enterprise Build of Drools queries support backward chaining, a reasoning technique where the system works backward from a goal:

query "ParentOf" (String parent, String child)
    Person(name == parent, children contains child)
    or
    (Person(name == parent, children contains $mid)
     and ParentOf($mid, child;))
end

This recursive query finds if a person is a parent of another person either directly or through multiple generations.

Unification

Queries support variable unification, allowing for more concise expressions:

query "FindAccountsForCustomer" (String customerId)
    $account: /accounts[customerId == customerId]
end

In this example, the parameter customerId is unified with the property access customerId in the pattern.

Live Queries

Live Queries extend the query capability with real-time updates when matching facts change:

// Execute live query
LiveQuery liveQuery = kieSession.openLiveQuery("FindPremiumCustomers",
    new Object[]{}, new ViewChangedEventListener() {
        @Override
        public void rowInserted(Row row) {
            Customer customer = (Customer) row.get("$customer");
            System.out.println("New premium customer: " + customer.getName());
        }

        @Override
        public void rowDeleted(Row row) {
            Customer customer = (Customer) row.get("$customer");
            System.out.println("No longer premium: " + customer.getName());
        }

        @Override
        public void rowUpdated(Row oldRow, Row newRow) {
            // Handle updates
        }
    });

// When done with live query
liveQuery.close();

Live queries are particularly useful for:

Building reactive user interfaces
Maintaining real-time dashboards
Monitoring systems

Query-based Rules

Queries can be used within rule conditions to create more modular and reusable rule structures:

rule "Apply Premium Discount"
when
    FindPremiumCustomers($customer;)
    $order: Order(customerId == $customer.id, status == "new")
then
    $order.applyDiscount(0.1);
    update($order);
end

This approach allows for separating the condition logic (in queries) from the action logic (in rules).

Best Practices for Queries

Name Queries Appropriately: Use clear, descriptive names that indicate what the query retrieves.
Parameterize When Possible: Use parameters to make queries more reusable.
Keep Queries Focused: Design queries with a single, clear purpose.
Consider Performance: Be mindful of the complexity of query patterns, especially with large fact bases.
Manage Variable Binding: Be consistent with variable naming to avoid confusion.
Document Complex Queries: Add comments to explain the purpose and logic of complex queries.
Test Queries Individually: Create dedicated tests for each query to verify they return the expected results.

Troubleshooting Queries

Issue	Possible Cause	Solution
Empty query results	No facts match the conditions	Verify conditions and check if facts have been inserted
ClassCastException	Incorrect variable types in results	Check that the retrieved objects match the expected types
Query not found	Incorrect query name or package	Verify the query name and package declaration
Parameter type mismatch	Passing parameters of wrong type	Check parameter types in the query definition
Unexpected infinite recursion	Recursive query without proper base case	Ensure recursive queries have termination conditions

Query Performance Considerations

Indexing: Aletyx Enterprise Build of Drools automatically creates indexes for some constraints. Position the most selective constraints first in patterns.
Query Complexity: Complex queries with many joins can be expensive. Consider breaking them into simpler queries if possible.
Data Volume: Be cautious with queries that might return large result sets.
Live Queries: Live queries have additional overhead. Use them only when necessary.

Integration with External Systems

Queries provide an excellent interface for integrating Aletyx Enterprise Build of Drools with external systems:

Databases: Use queries to retrieve rule engine data for storage
APIs: Expose queries as service endpoints
User Interfaces: Power UI components with query results
Reporting Systems: Feed business intelligence tools with query data

Conclusion

The Query system in Aletyx Enterprise Build of Drools provides a powerful mechanism for extracting and analyzing data from the rule engine. When combined with Rule Units, queries become even more effective as they can be scoped to specific domains and data sources. Mastering queries allows developers to build more interactive, dynamic, and data-driven rule-based applications. Whether used for simple data retrieval or complex pattern matching with backward chaining, queries enhance the flexibility and utility of the Aletyx Enterprise Build of Drools rule engine.

Home

Get started

AI Assistant

Core Concepts

Decision Control

Decision Control Tower

Drools

KIE Server

Guides

Platform

Deployment

Reference

FAQ

Query System in Aletyx Enterprise Build of Drools

Introduction to Drools Queries

Traditional Queries vs. Rule Unit Queries

Traditional Queries

Rule Unit Queries

Query Definition

Basic Structure

Pattern Matching in Queries

Parameters in Queries

Executing Queries

In Traditional Drools Applications

In Rule Unit Applications

Advanced Query Features

Nested Queries

Backward Chaining

Unification

Live Queries

Query-based Rules

Best Practices for Queries

Troubleshooting Queries

Query Performance Considerations

Integration with External Systems

Conclusion

​Introduction to Drools Queries

​Traditional Queries vs. Rule Unit Queries

​Traditional Queries

​Rule Unit Queries

​Query Definition

​Basic Structure

​Pattern Matching in Queries

​Parameters in Queries

​Executing Queries

​In Traditional Drools Applications

​In Rule Unit Applications

​Advanced Query Features

​Nested Queries

​Backward Chaining

​Unification

​Live Queries

​Query-based Rules

​Best Practices for Queries

​Troubleshooting Queries

​Query Performance Considerations

​Integration with External Systems

​Conclusion

Introduction to Drools Queries

Traditional Queries vs. Rule Unit Queries

Traditional Queries

Rule Unit Queries

Query Definition

Basic Structure

Pattern Matching in Queries

Parameters in Queries

Executing Queries

In Traditional Drools Applications

In Rule Unit Applications

Advanced Query Features

Nested Queries

Backward Chaining

Unification

Live Queries

Query-based Rules

Best Practices for Queries

Troubleshooting Queries

Query Performance Considerations

Integration with External Systems

Conclusion