Ignorar linhas anteriores no UNION com base na chave

Eu tenho uma tabela com prazos de entrega de tarefas, que é baseada no departamento de tarefas (apenas), categoria de tarefas (apenas) ou tarefa e categoria do produto. Essas regras podem ser aplicadas diretamente a um cliente, um grupo de clientes ou a todos os clientes.

Exemplo:

• Cliente 01 + Departamento: Design + Categoria: Criar Logo = 24 horas
• Cliente 01 + Departamento: Design + Qualquer Categoria = 48 horas Todos
• Clientes + Departamento: Design + Qualquer Categoria = 72 horas Todos
• Clientes + Qualquer Categoria + Categoria: Criar Logo = 48 horas

Esta tabela tem a estrutura abaixo:

|rule_id|customer_id|departament_id|category_id|deadline|

SQL para criar tabelas:

CREATE TABLE public.test_customer (
    id INTEGER NOT NULL,
    name varchar(50) NOT NULL
);
CREATE TABLE public.test_departament (
    id INTEGER NOT NULL,
    name varchar(50) NOT NULL
);
CREATE TABLE public.test_category (
    id INTEGER NOT NULL,
    name varchar(50) NOT NULL
);
CREATE TABLE public.test_task (
    id INTEGER NOT NULL,
    customer_id INTEGER,
    departament_id INTEGER,
    category_id INTEGER,
    description varchar(50) NOT NULL
);
CREATE TABLE public.test_time_rule (
    id INTEGER NOT NULL,
    customer_id INTEGER,
    departament_id INTEGER,
    category_id INTEGER,
    time_deadline INTEGER
);

INSERT INTO public.test_customer (id, name) VALUES(1, 'Customer with Rule');
INSERT INTO public.test_customer (id, name) VALUES(2, 'Customer without Rule');

INSERT INTO public.test_departament (id, name) VALUES(1, 'Front-End');
INSERT INTO public.test_departament (id, name) VALUES(2, 'Back-End');
INSERT INTO public.test_departament (id, name) VALUES(3, 'DBA');

INSERT INTO public.test_departament (id, name) VALUES(1, 'Design');
INSERT INTO public.test_departament (id, name) VALUES(2, 'Create HTML Pages');
INSERT INTO public.test_departament (id, name) VALUES(3, 'Create Tables');
INSERT INTO public.test_departament (id, name) VALUES(4, 'Fix Bugs');

INSERT INTO public.test_task (id, customer_id, departament_id, category_id, description) VALUES(1, 1, 1, 1, 'Create New Form Design');
INSERT INTO public.test_task (id, customer_id, departament_id, category_id, description) VALUES(2, 1, 2, 4, 'Fix Bug on Customer Table');
INSERT INTO public.test_task (id, customer_id, departament_id, category_id, description) VALUES(2, 1, 3, 3, 'Create City Table');
INSERT INTO public.test_task (id, customer_id, departament_id, category_id, description) VALUES(3, 2, 1, 1, 'Create New Form Design');
INSERT INTO public.test_task (id, customer_id, departament_id, category_id, description) VALUES(4, 2, 3, 3, 'Create City Table');

INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(1, NULL, NULL, NULL, 20); --All Customers
INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(2, NULL, 1, NULL, 30); --All Customers + Departament
INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(3, NULL, NULL, 1, 40); --All Customers + Category
INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(4, NULL, 3, 3, 50); --All Customers + Departament + Category

INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(5, 1, NULL, NULL, 20); --Customers With Rule
INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(6, 1, 1, NULL, 30); --Customers + Departament
INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(7, 1, NULL, 1, 40); --Customers + Category
INSERT INTO public.test_time_rule (id, customer_id, departament_id, category_id, time_deadline) VALUES(8, 1, 3, 3, 50); --Customers + Departament + Category

Estou tentando criar um SELECTque liste todas as tarefas e mostre qual regra se aplica a ele. Para isso estou utilizando os níveis de aplicação da regra, que são:

1 - Regra com cliente + departamento + categoria (ID de cliente definido)

2 - Regra com grupo de clientes + departamento + categoria (ID do grupo de clientes definido)

3 - Regra geral (todos os clientes) + departamento + categoria (sem cliente definido e ID do grupo de clientes)

4 - Regra com cliente + departamento

Eu resumi as regras, mas elas chegam a 16 regras, e criei uma view para cada regra com as tarefas que correspondem aos seus parâmetros.

Com base nessas regras criei um SELECTcom vários UNION ALL(motivos de performance), mas como a linha pode ser repetida devido a um select anterior, acabei adicionando NOT EXISTScondições para evitar esse problema.

Exemplo:

 SELECT
    1 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NOT NULL
    AND ttr.departament_id IS NOT NULL
    AND ttr.category_id IS NOT NULL
    AND tt.customer_id = ttr.customer_id
    AND tt.departament_id = ttr.departament_id
    AND tt.category_id = ttr.category_id
UNION
SELECT
    2 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NULL
    AND ttr.departament_id IS NOT NULL
    AND ttr.category_id IS NOT NULL
    AND tt.departament_id = ttr.departament_id
    AND tt.category_id = ttr.category_id
UNION
SELECT
    3 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NOT NULL
    AND ttr.departament_id IS NOT NULL
    AND ttr.category_id IS NULL
    AND tt.customer_id = ttr.customer_id
    AND tt.departament_id = ttr.departament_id
UNION
SELECT
    4 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NULL
    AND ttr.departament_id IS NOT NULL
    AND ttr.category_id IS NULL
    AND tt.departament_id = ttr.departament_id
UNION
SELECT
    5 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NOT NULL
    AND ttr.departament_id IS NULL
    AND ttr.category_id IS NOT NULL
    AND tt.customer_id = ttr.customer_id
    AND tt.category_id = ttr.category_id
UNION
SELECT
    6 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NULL
    AND ttr.departament_id IS NULL
    AND ttr.category_id IS NOT NULL
    AND tt.category_id = ttr.category_id
UNION
SELECT
    7 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NOT NULL
    AND ttr.departament_id IS NULL
    AND ttr.category_id IS NULL
    AND tt.customer_id = ttr.customer_id
UNION
SELECT
    8 as level,
    ttr.id as rule_id,
    ttr.customer_id as rule_customer_id,
    tt.customer_id as task_customer_id,
    ttr.departament_id as rule_departament_id,
    tt.departament_id as task_departament_id,
    ttr.category_id as rule_category_id,
    tt.category_id as task_category_id,
    tt.id as task_id,
    tt.description as task_description,
    ttr.time_deadline
FROM
    test_time_rule ttr,
    test_task tt
WHERE
    ttr.customer_id IS NULL
    AND ttr.departament_id IS NULL
    AND ttr.category_id IS NULL
ORDER BY
    level;

Como você deve ter notado a cada nível que ele entra, todas as visualizações anteriores são adicionadas, o que gera um SQL grande e lento.

O meu bom senso diz-me que esta não é a melhor forma de fazer esta consulta.

Considerei algo que ignora uma linha com base no ID da tarefa já exibido em uma união anterior. Existe uma maneira melhor de criar este SELECT?

Observação: não posso adicionar o ID da regra usada na tarefa, pois essa regra pode ser alterada, excluída ou até adicionada no futuro a uma regra que melhor corresponda à tarefa.