Background
The underlying structure of CDWPG is built on Greenplum 6. The PostgreSQL kernel is v9.4, and its
INSERT .. ON CONFLICT feature cannot be well supported for now. This document describes how to use a PostgreSQL rule to UPSERT, as it requires a different method.Rule Overview
The PostgreSQL rule system allows one to define an alternative action to be performed on insertions, updates, or deletions in database tables. Roughly speaking, a rule causes additional commands to be executed when a given command on a given table is executed. Alternatively, an
INSTEAD rule can replace a given command by another, or cause a command not to be executed at all. Rules are used to implement table views as well. A rule is really a command transformation mechanism, or command macro. The transformation happens before the execution of the command starts.UPSERT Rule
To implement an
UPSERT operation, you need a rule that determines whether a corresponding record already exists during INSERT, allows for UPDATE if yes, and allows for INSERT if no.The following database instance is used as an example:
Create a test database.
CREATE TABLE my_test(id integer,num1 integer,num2 decimal,str1 varchar(20),str2 text,PRIMARY KEY(id)) distributed by (id);
Then, add a rule to the table.
create rule r1 as on insert to my_test where exists (select 1 from e t1 where t1.id=NEW.id limit 1) do instead update my_test set num1=NEW.num1,num2=NEW.num2,str1=NEW.str1,str2=NEW.str2 where id=NEW.id;
This rule is for the
INSERT operation. If the id in the new INSERT statement already exists, the original data will be updated with the value inside the new INSERT, i.e., the NEW.XXX that can be seen after the operation. In this case, no errors will be reported due to the primary key constraint, and the UPDATE operation will be performed.\\d my_testTable "public.my_test"Column | Type | Collation | Nullable | Default--------+-----------------------+-----------+----------+-------------------------------------id | integer | | not null | nextval('my_test_id_seq'::regclass)num1 | integer | | |num2 | numeric | | |str1 | character varying(20) | | |str2 | text | | |Indexes:"my_test_pkey" PRIMARY KEY, btree (id)Rules:r1 ASON INSERT TO my_testWHERE (EXISTS ( SELECT 1FROM my_test my_test_1WHERE my_test_1.id = new.idLIMIT 1)) DO INSTEAD UPDATE my_test SET num1 = new.num1, num2 = new.num2, str1 = new.str1, str2 = new.str2
Notes
The rule is subject to the following limits:
1. Bulk insert may not be proper. If the statement does not set a unique constraint or primary key constraint, duplicate data may be generated during bulk insert. Therefore, you should avoid using bulk insert. When using it, avoid determining whether the fields of
UPSERT are not duplicated or add unique constraints to the fields that need to be determined. As shown in the following example, if there is no primary key constraint on the id, there may be duplicate data after execution.insert into my_test (id,num1,num2,str1,str2)values(1,2,1.0,'111','555'),(1,3,2.0,'111','666');
2. The rule does not support the
COPY statement, which may also cause duplicate data just like bulk insert.3. When you set the
UPDATE rule, if the rule usage is configured, but the INSERT statement does not pass in the num1 and num2 fields, these two fields will be null after UPDATE, resulting in the loss of the original data.update my_test set num1=NEW.num1,num2=NEW.num2,str1=NEW.str1,str2=NEW.str2
Yes
No
Was this page helpful?