I have a function (audit.create_audit_table()) that accepts an array of table names. It creates a single function audit.if_modified_func() and then loops through each table name and creates an audit table and applies a trigger to the main table. The function compiles and is created with no errors. When I run the function
select audit.create_audit_table(ARRAY['organization'])
I keep getting the following error and I am not sure why because I thought that TG_TABLE_NAME is a automatic variable which will give me access to the current table that is executing audit.if_modified_func()
ERROR:
ERROR: column "tg_table_name" does not exist
LINE 3: audit_row audit.' || quote_ident(TG_TABLE_NAME::TEXT)||';
^
Here's the function:
CREATE OR REPLACE FUNCTION audit.create_audit_table(table_names character varying[])
RETURNS character varying AS
$BODY$
DECLARE
table_name varchar;
i int;
BEGIN
EXECUTE 'CREATE OR REPLACE FUNCTION audit.if_modified_func() RETURNS TRIGGER AS $$
DECLARE
audit_row audit.' || quote_ident(TG_TABLE_NAME::TEXT)||';
include_values boolean;
log_diffs boolean;
h_old hstore;
h_new hstore;
excluded_cols text[] = ARRAY[]::text[];
BEGIN
IF TG_WHEN <> ''AFTER'' THEN
RAISE EXCEPTION ''audit.if_modified_func() may only run as an AFTER trigger'';
END IF;
audit_row = ROW(
nextval(''audit.'|| quote_ident(TG_TABLE_NAME::text) ||'_event_id_seq''), -- event_id
TG_TABLE_SCHEMA::text, -- schema_name
TG_TABLE_NAME::text, -- table_name
TG_RELID, -- relation OID for much quicker searches
session_user::text, -- session_user_name
current_timestamp, -- action_tstamp_tx
statement_timestamp(), -- action_tstamp_stm
clock_timestamp(), -- action_tstamp_clk
txid_current(), -- transaction ID
current_setting(''application_name''), -- client application
inet_client_addr(), -- client_addr
inet_client_port(), -- client_port
current_query(), -- top-level query or queries (if multistatement) from client
substring(TG_OP,1,1), -- action
NULL, NULL, -- row_data, changed_fields
''f'' -- statement_only
);
IF NOT TG_ARGV[0]::boolean IS DISTINCT FROM ''f''::boolean THEN
audit_row.client_query = NULL;
END IF;
IF TG_ARGV[1] IS NOT NULL THEN
excluded_cols = TG_ARGV[1]::text[];
END IF;
IF (TG_OP = ''UPDATE'' AND TG_LEVEL = ''ROW'') THEN
audit_row.row_data = hstore(OLD.*) - excluded_cols;
audit_row.changed_fields = (hstore(NEW.*) - audit_row.row_data) - excluded_cols;
IF audit_row.changed_fields = hstore('''') THEN
-- All changed fields are ignored. Skip this update.
RETURN NULL;
END IF;
ELSIF (TG_OP = ''DELETE'' AND TG_LEVEL = ''ROW'') THEN
audit_row.row_data = hstore(OLD.*) - excluded_cols;
ELSIF (TG_OP = ''INSERT'' AND TG_LEVEL = ''ROW'') THEN
audit_row.row_data = hstore(NEW.*) - excluded_cols;
ELSIF (TG_LEVEL = ''STATEMENT'' AND TG_OP IN (''INSERT'',''UPDATE'',''DELETE'',''TRUNCATE'')) THEN
audit_row.statement_only = ''t'';
ELSE
RAISE EXCEPTION ''[audit.if_modified_func] - Trigger func added as trigger for unhandled case: %%, %%'',TG_OP, TG_LEVEL;
RETURN NULL;
END IF;
INSERT INTO audit.'|| quote_ident(TG_TABLE_NAME::TEXT) ||' VALUES (audit_row.*);
RETURN null;
END;
$$
LANGUAGE plpgsql;
ALTER FUNCTION audit.if_modified_func()
OWNER TO postgres;';
FOR i in 1..array_upper(table_names, 1) LOOP
EXECUTE format('
DROP TABLE IF EXISTS audit.%1$s;
CREATE TABLE audit.%1$s (
event_id bigserial primary key,
schema_name text not null,
table_name text not null,
relid oid not null,
session_user_name text,
action_tstamp_tx TIMESTAMP WITH TIME ZONE NOT NULL,
action_tstamp_stm TIMESTAMP WITH TIME ZONE NOT NULL,
action_tstamp_clk TIMESTAMP WITH TIME ZONE NOT NULL,
transaction_id bigint,
application_name text,
client_addr inet,
client_port integer,
client_query text,
action TEXT NOT NULL CHECK (action IN (''I'',''D'',''U'', ''T'')),
row_data hstore,
changed_fields hstore,
statement_only boolean not null
);
REVOKE ALL ON audit.%1$s FROM public;
COMMENT ON TABLE audit.%1$s IS ''History of auditable actions on audited tables, from audit.if_modified_func()'';
COMMENT ON COLUMN audit.%1$s.event_id IS ''Unique identifier for each auditable event'';
COMMENT ON COLUMN audit.%1$s.schema_name IS ''Database schema audited table for this event is in'';
COMMENT ON COLUMN audit.%1$s.table_name IS ''Non-schema-qualified table name of table event occured in'';
COMMENT ON COLUMN audit.%1$s.relid IS ''Table OID. Changes with drop/create. Get with ''''tablename''''::regclass'';
COMMENT ON COLUMN audit.%1$s.session_user_name IS ''Login / session user whose statement caused the audited event'';
COMMENT ON COLUMN audit.%1$s.action_tstamp_tx IS ''Transaction start timestamp for tx in which audited event occurred'';
COMMENT ON COLUMN audit.%1$s.action_tstamp_stm IS ''Statement start timestamp for tx in which audited event occurred'';
COMMENT ON COLUMN audit.%1$s.action_tstamp_clk IS ''Wall clock time at which audited event''''s trigger call occurred'';
COMMENT ON COLUMN audit.%1$s.transaction_id IS ''Identifier of transaction that made the change. May wrap, but unique paired with action_tstamp_tx.'';
COMMENT ON COLUMN audit.%1$s.client_addr IS ''IP address of client that issued query. Null for unix domain socket.'';
COMMENT ON COLUMN audit.%1$s.client_port IS ''Remote peer IP port address of client that issued query. Undefined for unix socket.'';
COMMENT ON COLUMN audit.%1$s.client_query IS ''Top-level query that caused this auditable event. May be more than one statement.'';
COMMENT ON COLUMN audit.%1$s.application_name IS ''Application name set when this audit event occurred. Can be changed in-session by client.'';
COMMENT ON COLUMN audit.%1$s.action IS ''Action type; I = insert, D = delete, U = update, T = truncate'';
COMMENT ON COLUMN audit.%1$s.row_data IS ''Record value. Null for statement-level trigger. For INSERT this is the new tuple. For DELETE and UPDATE it is the old tuple.'';
COMMENT ON COLUMN audit.%1$s.changed_fields IS ''New values of fields changed by UPDATE. Null except for row-level UPDATE events.'';
COMMENT ON COLUMN audit.%1$s.statement_only IS ''''''t'''' if audit event is from an FOR EACH STATEMENT trigger, ''''f'''' for FOR EACH ROW'';
CREATE INDEX %1$s_relid_idx ON audit.%1$s(relid);
CREATE INDEX %1$s_action_tstamp_tx_stm_idx ON audit.%1$s(action_tstamp_stm);
CREATE INDEX %1$s_action_idx ON audit.%1$s(action);
', table_names[i]);
EXECUTE format('
DROP TRIGGER IF EXISTS audit_trigger_row ON %1$s;
CREATE TRIGGER audit_trigger_row
AFTER INSERT OR UPDATE OR DELETE
ON public.%1$s
FOR EACH ROW
EXECUTE PROCEDURE audit.if_modified_func();', table_names[i]);
EXECUTE format('
DROP TRIGGER IF EXISTS audit_trigger_stm ON %1$s;
CREATE TRIGGER audit_trigger_stm
AFTER TRUNCATE
ON public.%1$s
FOR EACH STATEMENT
EXECUTE PROCEDURE audit.if_modified_func();', table_names[i]);
END LOOP;
RETURN 'SUCCESS';
END;
$BODY$
LANGUAGE plpgsql;
ALTER FUNCTION audit.create_audit_table(character varying[])
OWNER TO postgres;
UPDATE 03/31:
Ok, so I created the if_modified_func() function without the dynamic sql and I declared the audit_row as audit_row RECORD; I am not sure about the part of "needing a cast upon inserting the values". I am also not sure if this is the correct way to do the insert
EXECUTE format($string$INSERT INTO audit.%1$s VALUES (audit_row.*);$string$, TG_TABLE_NAME::text);
I am now getting this error when I run select audit.create_audit_table(ARRAY['organization'])
ERROR:
ERROR: record "audit_row" has no field "row_data"
CONTEXT: PL/pgSQL function audit.if_modified_func() line 42 at assignment
Here's the updated function:
CREATE OR REPLACE FUNCTION audit.if_modified_func() RETURNS TRIGGER AS $$
DECLARE
audit_row RECORD;
include_values boolean;
log_diffs boolean;
h_old hstore;
h_new hstore;
excluded_cols text[] = ARRAY[]::text[];
BEGIN
IF TG_WHEN <> 'AFTER' THEN
RAISE EXCEPTION 'audit.if_modified_func() may only run as an AFTER trigger';
END IF;
audit_row = ROW(
nextval(format('audit.%1$s_event_id_seq',TG_TABLE_NAME::text)), -- event_id
TG_TABLE_SCHEMA::text, -- schema_name
TG_TABLE_NAME::text, -- table_name
TG_RELID, -- relation OID for much quicker searches
session_user::text, -- session_user_name
current_timestamp, -- action_tstamp_tx
statement_timestamp(), -- action_tstamp_stm
clock_timestamp(), -- action_tstamp_clk
txid_current(), -- transaction ID
current_setting('application_name'), -- client application
inet_client_addr(), -- client_addr
inet_client_port(), -- client_port
current_query(), -- top-level query or queries (if multistatement) from client
substring(TG_OP,1,1), -- action
NULL, NULL, -- row_data, changed_fields
'f' -- statement_only
);
IF NOT TG_ARGV[0]::boolean IS DISTINCT FROM 'f'::boolean THEN
audit_row.client_query = NULL;
END IF;
IF TG_ARGV[1] IS NOT NULL THEN
excluded_cols = TG_ARGV[1]::text[];
END IF;
IF (TG_OP = 'UPDATE' AND TG_LEVEL = 'ROW') THEN
audit_row.row_data = hstore(OLD.*) - excluded_cols;
audit_row.changed_fields = (hstore(NEW.*) - audit_row.row_data) - excluded_cols;
IF audit_row.changed_fields = hstore('') THEN
-- All changed fields are ignored. Skip this update.
RETURN NULL;
END IF;
ELSIF (TG_OP = 'DELETE' AND TG_LEVEL = 'ROW') THEN
audit_row.row_data = hstore(OLD.*) - excluded_cols;
ELSIF (TG_OP = 'INSERT' AND TG_LEVEL = 'ROW') THEN
audit_row.row_data = hstore(NEW.*) - excluded_cols;
ELSIF (TG_LEVEL = 'STATEMENT' AND TG_OP IN ('INSERT','UPDATE','DELETE','TRUNCATE')) THEN
audit_row.statement_only = 't';
ELSE
RAISE EXCEPTION '[audit.if_modified_func] - Trigger func added as trigger for unhandled case: %, %',TG_OP, TG_LEVEL;
RETURN NULL;
END IF;
EXECUTE format('INSERT INTO audit.%1$s VALUES (audit_row.*)', TG_TABLE_NAME::text);
RETURN null;
END;
$$
LANGUAGE plpgsql;
ALTER FUNCTION audit.if_modified_func()
OWNER TO postgres;
TG_TABLE_NAME is a special, trigger variable, which is only available inside trigger functions. Your create_audit_table() is not a trigger function.
Also, you constantly redefining your real trigger function (if_modified_func()), which "invalidates" any earlier created triggers.
Create your trigger function without the dynamic SQL magic (dynamic SQL will only need to insert values to these audit tables). Then, you can add your audit logic to a table with:
CREATE TRIGGER audit_trigger_row
AFTER INSERT OR UPDATE OR DELETE
ON public.<your_table_name>
FOR EACH ROW
EXECUTE PROCEDURE <your_audit_trigger_function_name>();
You can put this (but only this -- maybe with drop if exists) inside a function, to allow attaching this audit logic more easily.
Notes:
Inside the trigger function, you cannot use a %ROWTYPE variable (because you don't know the exact table. you only have its name). The solution is simple: just use the RECORD type instead (you will need a cast upon inserting the values though).
Don't use single quotes for such long strings. Use the $your_keyword$<string_value>$your_keyword$ format instead. With possibly the format() function instead of just concatenating values. Your code will be much more readable.
Edit: to utilize your RECORD variable, you should either:
Initialize it with a structure. You can do this in your case with f.ex.
SELECT nextval('audit.'|| quote_ident(TG_TABLE_NAME) || '_event_id_seq') AS event_id,
TG_TABLE_SCHEMA AS schema_name,
TG_TABLE_NAME AS table_name,
TG_RELID AS relid,
session_user AS session_user_name,
current_timestamp AS action_tstamp_tx,
statement_timestamp() AS action_tstamp_stm,
clock_timestamp() AS action_tstamp_clk,
txid_current() AS transaction_id,
current_setting('application_name') AS application_name,
inet_client_addr() AS client_addr,
inet_client_port() AS client_port,
current_query() AS client_query,
substring(TG_OP, 1, 1) AS action,
NULL::hstore AS row_data,
NULL::hstore AS changed_fields,
FALSE AS statement_only
INTO audit_row;
Use the predefined names of the ROW() constructor. The first column's name if f1, the second's is f2, etc.
audit_row.f15 = hstore(OLD.*) - excluded_cols;
After choosing one of the above methods, you should insert the row like:
EXECUTE format('INSERT INTO audit.%1$s VALUES (($1::text::audit.%1$s).*)', quote_ident(TG_TABLE_NAME)) USING audit_row;
Note: even the cast to text is required due to the fact that EXECUTE cannot know the actual structure of audit_row.
http://rextester.com/GUAJ1339
quote_ident(TG_TABLE_NAME::TEXT) will apply necessary actions to correcty quote the argument as relation name.
I would recommend using execute format('statement') instead of concatinations, eg:
t=# do $$ begin raise info '%',format('I am %I, now is %L',current_user,now()); end;$$;
INFO: I am postgres, now is '2017-03-30 07:33:53.579476+00'
DO
Instead of:
t=# do $$ begin raise info '%','I am '||quote_ident(current_user)||', now is '||quote_ident(now()::text); end;$$;
INFO: I am postgres, now is "2017-03-30 07:36:20.495887+00"
DO
Related
I came across a strange behavior (at least for me) with PostgreSQL and a trigger BEFORE UPDATE.
I have a table witch has an updated_at column witch is set by a BEFORE UPDATE trigger.
I need to add new columns to this table and set their values with an UPDATE query (not with DEFAULT).
It works just fine excepts when i do an UPDATE juste before adding those columns.
Here's an example :
ALTER TABLE my_schema.my_table ADD COLUMN new_column varchar(50);
UPDATE my_schema.my_table SET new_column = 'new_column_update' WHERE id = xxxxxx;
This script works fine.
But if i do an UPDATE before :
UPDATE my_schema.my_table SET other_column = 'other_column_update' WHERE id = xxxxxx; -- the TRIGGER is triggered
ALTER TABLE my_schema.my_table ADD COLUMN new_column varchar(50);
UPDATE my_schema.my_table SET new_column = 'new_column_update' WHERE id = xxxxxx; -- this UPDATE does't do anything
It doesn't works anymore.
After a few (a lot) hours, i found that the trigger BEFORE UPDATE is reponsible. But i can't find why.
I found a workaround by temporary disabling the trigger
ALTER TABLE my_table DISABLE TRIGGER update_date;
Here is a dbfiddle, just run it to see this behaviour :
dbfiddle
Here is the code in dbfiddle
CREATE TABLE my_table (
other_column varchar(50),
updated_at timestamp
);
CREATE OR REPLACE FUNCTION update_date()
RETURNS trigger
LANGUAGE plpgsql
COST 1
AS '
BEGIN
IF row(NEW.*) IS DISTINCT FROM row(OLD.*) THEN
NEW.updated_at = now();
RETURN NEW;
ELSE
RETURN OLD;
END IF;
END;
'
;
CREATE TRIGGER update_date BEFORE
UPDATE
ON
my_table FOR EACH ROW EXECUTE PROCEDURE update_date();
INSERT INTO my_table VALUES ('other_column_insert');
UPDATE my_table SET other_column = 'other_column_update';
ALTER TABLE my_table ADD COLUMN new_colum varchar(50);
UPDATE my_table SET new_colum = 'new_colum_update'; -- this UPDATE doesn't work because of the trigger BEFORE UPDATE
-- It is possible to make it works by disabling the trigger BEFORE the first UPDATE
-- ALTER TABLE my_table DISABLE TRIGGER update_date;
Have you ever encountered this behavior ?
It's something to do with the (unnecessary) wrapping of NEW/OLD with a ROW(...) constructor:
BEGIN
IF row(NEW.*) IS DISTINCT FROM row(OLD.*) THEN
-- IF NEW IS DISTINCT FROM OLD THEN
NEW.updated_at = now();
ELSE
RAISE EXCEPTION $$NOT DISTINCT: % / %$$, NEW, OLD;
END IF;
RETURN NEW;
END;
I've also moved the RETURN NEW to the end. If you try your version you should see the exceptions. If you replace it out with the commented-out one below then it works.
Now, as to why this is failing when you compare rows I'm not sure and it's too hot and late on a Friday afternoon where I am to figure it out I'm afraid.
I am going to say this is a caching problem. I modified the function to see what is going on:
DROP TABLE IF EXISTS my_table;
CREATE TABLE my_table (
other_column varchar(50),
updated_at timestamp
);
CREATE OR REPLACE FUNCTION public.update_date()
RETURNS trigger
LANGUAGE plpgsql
COST 1
AS $function$
BEGIN
RAISE NOTICE 'New row %', ROW(NEW.*);
RAISE NOTICE 'Old row%', ROW(OLD.*);
RAISE NOTICE 'New.* %', (NEW.*)::text;
RAISE NOTICE 'Old.* %', (OLD.*)::text;
IF NEW.* IS DISTINCT FROM OLD.* THEN
NEW.updated_at = now();
RETURN NEW;
ELSE
RETURN OLD;
END IF;
END;
$function$;
CREATE TRIGGER update_date BEFORE
UPDATE
ON
my_table FOR EACH ROW EXECUTE PROCEDURE update_date();
INSERT INTO my_table VALUES ('other_column_insert');
UPDATE my_table SET other_column = 'other_column_update';
NOTICE: New row (other_column_update,)
NOTICE: Old row(other_column_insert,)
NOTICE: New.* (other_column_update,)
NOTICE: Old.* (other_column_insert,)
ALTER TABLE my_table ADD COLUMN new_colum varchar(50);
UPDATE my_table SET new_colum = 'new_colum_update';
NOTICE: New row (other_column_update,"2022-08-12 10:38:54.815831")
NOTICE: Old row(other_column_update,"2022-08-12 10:38:54.815831")
NOTICE: New.* (other_column_update,"2022-08-12 10:38:54.815831",new_colum_update)
NOTICE: Old.* (other_column_update,"2022-08-12 10:38:54.815831",)
It has to do with the ROW(). Even doing ROW(NEW.*)::my_table or using EXECUTE to make the query dynamic and not use caching does not work.
I have been working on creating a store procedure that will select data from a table, do some modification to that data, and then I need to insert that modified data into the same table. Take an example my table name is student. My procedure looks like below:
create or replace procedure student_create(p_code varchar)
language plpgsql
as $$
declare
v_student public.student;
begin
select * into v_student from student where code = p_code and is_latest_version is true;
raise notice 'Value: %', v_student;
v_student.id = uuid_generate_v4();
v_student.version_created_at = now();
v_student.version_updated_at = v_student.version_created_at;
raise notice 'Value: %', v_student;
INSERT INTO public.student VALUES(v_student);
end;$$
I am getting errors while calling this procedure:
ERROR: column "id" is of type uuid but expression is of type hotel
LINE 1: INSERT INTO public.hotel VALUES(v_hotel)
I know I can make insert statements like I can get each value from the variable and set it like
INSERT INTO public.student VALUES(v_student.id, v_student.code, v_student.name);
But I don't want to do that because it will become tightly coupled and later if I add another column into the table then I need to add that column into this procedure as well.
Does anyone have idea how can I insert the declared type variable directly into table.
There is no table type, there is only row composite type. Check manual 43.3.4. Row Types.
use row type.
create or replace procedure student_create(p_code text)
language plpgsql
as $$
declare
v_student public.student
begin
for v_student in select * from student where code = p_code and is_latest_version is true
loop
v_student.id = uuid_generate_v4();
v_student.version_created_at = now();
v_student.version_updated_at = v_student.version_created_at;
v_student.is_latest_version = true;
v_student.code = p_code;
INSERT INTO student VALUES(v_student.*);
end loop;
end;$$;
call it: call student_create('hello');
3. use update clause directly.
create or replace procedure student_create_1(p_code text)
language plpgsql as $$
BEGIN
with a as ( select uuid_generate_v4() as id ,
now() as version_created_at,
now() as version_updated_at,
p_code as "code" from student
where code = p_code and is_latest_version is true)
INSERT INTO student(id, version_created_at, version_updated_at, code)
select a.id, a.version_created_at,a.version_updated_at,a."code" from a;
end
$$;
call it: call student_create_1('hello');
fiddle code: here
In Postgres, someone knows how to substitute the value of the variable in a NEW.variable in a trigger?
For instance, I have a variable with value order_code. I want to execute NEW.variable so that it's getting in fact NEW.order_code.
In detailed:
I have a function to obtain the primary key column of a table:
CREATE FUNCTION getPrimaryKey(_table_name VARCHAR(50))
RETURNS SETOF VARCHAR(50) AS $$
DECLARE
primary_key VARCHAR(50);
BEGIN
FOR primary_key IN SELECT a.attname
FROM pg_index i
JOIN pg_attribute a ON a.attrelid = i.indrelid
AND a.attnum = ANY(i.indkey)
WHERE i.indrelid = _table_name::regclass
AND i.indisprimary LOOP
RETURN NEXT primary_key;
END LOOP;
END;
$$ LANGUAGE plpgsql;
Then I have a trigger to collect some info when an INSERT is done in a table. The procedure in the trigger is called from several triggers from different tables. That's why it's so generic and I have this need.
What I want is to obtain the primary key of the object inserted.
CREATE FUNCTION logAudit()
RETURNS trigger AS $$
DECLARE primary_key VARCHAR(50);
BEGIN
primary_key := getprimarykey(TG_TABLE_NAME::VARCHAR(50));
INSERT INTO test VALUES (TG_TABLE_NAME);
INSERT INTO test VALUES (NEW.primary_key);
RETURN NULL;
END;
$$ LANGUAGE plpgsql;
CREATE TRIGGER audit_in_client
AFTER INSERT ON tb_client
FOR EACH STATEMENT EXECUTE PROCEDURE logAudit();
The NEW.primary_key is what is causing me issues. I expect primary_key to be the column name of the source table where the insert happened. What I want in NEW.primary_key is to actually use the value in the variable.
Here is the example of anonymous pl/pgsql block which doing something what you want:
do $$
declare
v pg_database = (pg_database) from pg_database where datname = 'template1';
fname text = 'datname';
n text;
begin
n := to_jsonb(v)->>fname;
raise info '%', n;
end $$;
Output:
INFO: template1
It is working example. In your trigger function it could be something like
declare
pk_name text;
pk_value text;
begin
pk_name := getprimarykey(TG_TABLE_NAME::VARCHAR(50));
pk_value := to_jsonb(NEW) ->> pk_name;
-- Do what you want with pk_value here
return null;
end $$;
I wanted to have an explanation on triggers of Postgres views.
To make clear what I want to ask, I'll give you a very simplified example of my case.
In this example we have two tables (table_a, table_b) that joined together make the view in the example (vw_table_ab).
In this example I will use trivial names and simple DDLs/DMLs.
-- TABLE table_a
CREATE TABLE table_a
(
id serial PRIMARY KEY,
timestamp_field timestamp DEFAULT now() NOT NULL,
boolean_field boolean DEFAULT FALSE NOT NULL
);
-- TABLE table_b
CREATE TABLE table_b
(
id serial PRIMARY KEY,
timestamp_field timestamp DEFAULT now() NOT NULL,
boolean_field boolean DEFAULT FALSE NOT NULL,
id_table_a integer NOT NULL,
CONSTRAINT "fk_table_a" FOREIGN KEY (id_table_a) REFERENCES table_a (id) ON DELETE CASCADE NOT DEFERRABLE,
CONSTRAINT "u_table_a" UNIQUE (id_table_a)
);
-- VIEW vw_table_ab
CREATE VIEW vw_table_ab AS (
SELECT a.timestamp_field AS timestamp_a,
a.boolean_field AS boolean_a,
b.timestamp_field AS timestamp_b,
b.boolean_field AS boolean_b
FROM table_a a
JOIN table_b b ON a.id = b.id_table_a
);
A trigger function on standard actions (INSERT, UPDATE and DELETE) is linked to this view through an INSTEAD OF trigger.
-- TRIGGER FUNCTION fn_trigger
CREATE FUNCTION fn_trigger() RETURNS trigger LANGUAGE plpgsql AS
$_$
DECLARE
sql TEXT;
BEGIN
sql = 'SELECT ' || TG_TABLE_NAME || '_' || lower(TG_OP) || '($1);';
IF TG_OP = 'INSERT' OR TG_OP = 'UPDATE' THEN
EXECUTE (sql) USING NEW;
RAISE NOTICE '%', sql;
RETURN NEW;
ELSE
EXECUTE (sql) USING OLD;
RAISE NOTICE '%', sql;
RETURN OLD;
END IF;
END;
$_$;
-- TRIGGER tr_table_ab
CREATE TRIGGER tr_table_ab
INSTEAD OF INSERT OR UPDATE OR DELETE ON vw_table_ab
FOR EACH ROW EXECUTE PROCEDURE fn_trigger();
The example I bring has a trigger called only on the insert action, and the function that is executed is this:
-- INSERT FUNCTION vw_table_ab_insert
CREATE FUNCTION vw_table_ab_insert(new vw_table_ab) RETURNS void LANGUAGE plpgsql AS
$_$
DECLARE
id_table_a integer;
BEGIN
INSERT INTO table_a (timestamp_field, boolean_field) VALUES (new.timestamp_a, new.boolean_a)
RETURNING id
INTO id_table_a;
INSERT INTO table_b (timestamp_field, boolean_field, id_table_a) VALUES (new.timestamp_a, new.boolean_b, id_table_a);
END;
$_$;
Now we can get to my problem. I make an insert on the view, and when the action is triggered, I get a "Not null violation" error becouse I have some NOT NULL constraints on table_a and table_b like in this case:
INSERT INTO vw_table_ab (timestamp_a, boolean_a, timestamp_b, boolean_b) VALUES (now(), NULL, now(), NULL);
Suppose that the previous statement is generated through a programming language framework and I don't want to handle this case in backend code, but I want handle this case in PostgreSQL in the insert function vw_table_ab_insert. So at this point my problem is bound to the new parameter of the function because I have fields of the view that are NULL. But these fields have a DEFAULT value in the definition of the base table, and I want to use that.
...
timestamp_field timestamp DEFAULT now() NOT NULL,
boolean_field boolean DEFAULT FALSE NOT NULL
...
My question:
How can I manage the NULL values in trigger of the views using the DEFAULT defined in the tables?
Initially I thought of putting IF ... THEN ... inside the function and override null values with DEFAULT expression but I do not really like that.
For example, the function would become like this:
CREATE FUNCTION vw_table_ab_insert(new vw_table_ab) RETURNS void LANGUAGE plpgsql AS
$_$
DECLARE
id_table_a integer;
BEGIN
IF new.timestamp_a IS NULL THEN
new.timestamp_a = DEFAULT;
END IF;
IF new.boolean_a IS NULL THEN
new.boolean_a = DEFAULT;
END IF;
IF new.timestamp_b IS NULL THEN
new.timestamp_b = DEFAULT;
END IF;
IF new.boolean_b IS NULL THEN
new.boolean_b = DEFAULT;
END IF;
INSERT INTO table_a (timestamp_field, boolean_field)
VALUES (new.timestamp_a, new.boolean_a)
RETURNING id
INTO id_table_a;
INSERT INTO table_b (timestamp_field, boolean_field, id_table_a)
VALUES (new.timestamp_a, new.boolean_b, id_table_a);
END;
$_$;
Someone can help me? Is there another method for handling this case?
The easiest way would be to use ALTER VIEW ... ALTER col SET DEFAULT to define default values on the view that are the same as the default values on the base table.
Then instead of inserting explicit NULLs, omit the columns from the INSERT statement or insert DEFAULT explicitly. Your resulting view will behave just like a real table.
I want to keep all changes of my tables. I have a working solution for making a trigger per table, but it seems silly to copy the code foreach table. Is there any way to create a single trigger function that does this?
Example of my working per-table trigger (including table definitions):
CREATE TABLE departments (
id bigserial Primary Key,
name varchar not null,
created bigint not null default date_part('epoch', NOW()),
created_by bigint references Employees (id) not null
);
create table Departments_hist ("action" varchar not null, change_date bigint not null, rev bigserial not null, like Departments);
CREATE OR REPLACE FUNCTION add_to_history_Departments() RETURNS TRIGGER AS $$
BEGIN
IF(TG_OP='INSERT' OR TG_OP='UPDATE') THEN
INSERT INTO Departments_hist values (TG_OP,date_part('epoch', NOW()),DEFAULT,NEW.*);
END IF;
IF (TG_OP='DELETE') THEN
INSERT INTO Departments_hist values (TG_OP,date_part('epoch', NOW()),DEFAULT,OLD.*);
END IF;
RETURN NEW;
END;
$$ LANGUAGE plpgsql;
CREATE TRIGGER trigger_history_Departments AFTER INSERT OR UPDATE OR DELETE ON Departments FOR EACH ROW EXECUTE PROCEDURE add_to_history_Departments();
I've tried to make it multi-table by concatenating '_hist' to TG_TABLE_NAME:
CREATE OR REPLACE FUNCTION add_to_hist_table() RETURNS TRIGGER AS $$
DECLARE
histTable text :=TG_TABLE_NAME || '_hist';
BEGIN
IF (TG_OP='INSERT' OR TG_OP='UPDATE') THEN
INSERT INTO histTable values (TG_OP,date_part('epoch', NOW()),DEFAULT,NEW.*);
ELSIF TG_OP='DELETE' THEN
INSERT INTO histTable values (TG_OP,date_part('epoch', NOW()),DEFAULT,OLD.*);
END IF;
RETURN null; --ignored since it is an AFTER triggger.
END;
$$ LANGUAGE plpgsql;
But I get an error:
ERROR: syntax error at or near "$1"
LINE 1: INSERT INTO $1 values ( $2 ,date_part('epoch', NOW()),DEFA...
^
QUERY: INSERT INTO $1 values ( $2 ,date_part('epoch', NOW()),DEFAULT, $3 .*)
CONTEXT: SQL statement in PL/PgSQL function "add_to_hist_table" near line 5
I guess it is a problem with variable substitution ( http://www.postgresql.org/docs/8.4/static/plpgsql-implementation.html ).
How can this functionality be achieved?
PS. I'm using postgresql 8.4 but will likely upgrade to 9.3 soon.
I found a solution on this "related question" https://stackoverflow.com/a/1997417/844731
I didn't think of doing 'EXECUTE USING' with NEW and OLD. So now a working solution is:
CREATE OR REPLACE FUNCTION add_to_hist_table() RETURNS TRIGGER AS $$
BEGIN
IF (TG_OP='INSERT' OR TG_OP='UPDATE') THEN
execute 'INSERT INTO '|| TG_TABLE_NAME ||'_hist values (''' || TG_OP || ''',date_part(''epoch'', NOW()),DEFAULT,$1.*)' using NEW;
ELSIF TG_OP='DELETE' THEN
execute 'INSERT INTO '|| TG_TABLE_NAME ||'_hist values (''' || TG_OP || ''',date_part(''epoch'', NOW()),DEFAULT,$1.*)' using OLD;
END IF;
RETURN null; --ignored since it is an AFTER triggger.
END;
$$ LANGUAGE plpgsql;
#Pascal_dher, somebody can create table with name with containing attacker code. Due Postgresql this probably dosnt do some really bad, only failed queries. But if your trigger will be more complex, then impacts can be much more worse.