postgres: How to generically make a column immutable? - postgresql

Here's the problem.
create table customer (
customer_id int generated by default as identity (start with 100) primary key
);
create table cart (
cart_id int generated by default as identity (start with 100) primary key
);
I want to protect customer_id and cart_id from updating generically once they are inserted. How?
UPD: While I was writing the question I found the answer to my original question. Here it is:
create table cart (
cart_id int generated by default as identity (start with 100) primary key,
name text not null,
at timestamp with time zone
);
create or replace function table_update_guard() returns trigger
language plpgsql immutable parallel safe cost 1 as $body$
begin
raise exception
'trigger %: updating is prohibited for %',
tg_name, tg_argv[0]
using errcode = 'restrict_violation';
return null;
end;
$body$;
create or replace trigger cart_update_guard
before update of cart_id, name on cart for each row
-- NOTE: the WHEN clause below is optional
when (
old.cart_id is distinct from new.cart_id
or old.name is distinct from new.name
)
execute function table_update_guard('cart_id, name');
> insert into cart (cart_id, name) values (0, 'prado');
INSERT 0 1
> update cart set cart_id = -1 where cart_id = 0;
ERROR: trigger cart_update_guard: updating is prohibited for cart_id, name
CONTEXT: PL/pgSQL function table_update_guard() line 3 at RAISE
> update cart set name = 'nasa' where cart_id = 0;
ERROR: trigger cart_update_guard: updating is prohibited for cart_id, name
CONTEXT: PL/pgSQL function table_update_guard() line 3 at RAISE
> update cart set at = now() where cart_id = 0;
UPDATE 1
The WHEN clause was suggested by Belayer in his answer. The full explanation is in my research. Additionally I examined the approach with playing with privileges. NOTE: Some people say that triggers like here are performance killers. They are wrong. How do you think postgres implements constraints internally? — Using implicit triggers like defined here.

If I understand correctly you want to prevent any user from modifying the the table id once it is established and to have a generic function produce the exception, while still allowing other updates. You can accomplish this by modifying the trigger rather than the function. Specify the WHEN predicate on the trigger itself. For the cart table then:
create or replace trigger cart_id_guard
before update of cart_id
on cart for each row
when (old.cart_id is distinct from new.cart_id)
execute function surrogate_id_guard('cart_id');
The for the customer table the trigger becomes:
create or replace trigger customer_id_guard
before update of customer_id
on customer for each row
when (old.customer_id is distinct from new.customer_id)
execute function surrogate_id_guard('customer_id');
The trigger function itself does not change. (demo here)

TL;DR
What did I try? Revoking UPDATE privilege doesn't work.
# \c danissimo danissimo
You are now connected to database "danissimo" as user "danissimo".
> revoke update (customer_id) on customer from danissimo;
REVOKE
> insert into customer (customer_id) values (0);
INSERT 0 1
> update customer set customer_id = 0 where customer_id = 0;
UPDATE 1
> update customer set customer_id = -1 where customer_id = 0;
UPDATE 1
Okay, let's put a guard on it.
create or replace function customer_id_guard() returns trigger
language plpgsql as $body$
begin
if old.customer_id != new.customer_id then
raise exception
'trigger %: updating is prohibited for %',
tg_name, 'customer_id' using
errcode = 'restrict_violation';
end if;
return new;
end;
$body$;
create or replace trigger customer_id_guard
after update on customer for each row
execute function customer_id_guard();
Now let's give them some work.
> update customer set customer_id = -1 where customer_id = -1;
UPDATE 1
Right, I didn't change the value. What about this:
> update customer set customer_id = 0 where customer_id = -1;
ERROR: trigger customer_id_guard: updating is prohibited for customer_id
CONTEXT: PL/pgSQL function customer_id_guard() line 4 at RAISE
Yeah, here it goes. Good, let's protect cart_id as well. I don't want to copy–paste trigger functions, so I let's try to generalize it:
create or replace function generated_id_guard() returns trigger
language plpgsql as $body$
declare
id_col_name text := tg_argv[0];
equal boolean;
begin
execute format('old.%1$I = new.%1$I', id_col_name) into equal;
if not equal then
raise exception
'trigger %: updating is prohibited for %',
tg_name, id_col_name using
errcode = 'restrict_violation';
end if;
return new;
end;
$body$;
create or replace trigger cart_id_guard
after update on cart for each row
execute function generated_id_guard('cart_id');
As you might notice I pass the column name to the trigger function and generate an expression and put the result of that expression into equal which then test.
> insert into cart (cart_id) values (0);
INSERT 0 1
> update cart set cart_id = 0 where cart_id = 0;
ERROR: syntax error at or near "old"
LINE 1: old.cart_id = new.cart_id
^
QUERY: old.cart_id = new.cart_id
CONTEXT: PL/pgSQL function generated_id_guard() line 6 at EXECUTE
Hmmm... He's right, what the dangling old.cart_id = new.cart_id? What if I write
execute format('select old.%1$I = new.%1$I', id_col_name) into equal;
> update cart set cart_id = 0 where cart_id = 0;
ERROR: missing FROM-clause entry for table "old"
LINE 1: select old.cart_id = new.cart_id
^
QUERY: select old.cart_id = new.cart_id
CONTEXT: PL/pgSQL function generated_id_guard() line 6 at EXECUTE
Right, right... What if I write
declare
id_old int;
id_new int;
begin
execute format('select %I from old', id_col_name) into id_old;
execute format('select %I from new', id_col_name) into id_new;
if id_old != id_new then
> update cart set cart_id = 0 where cart_id = 0;
ERROR: relation "old" does not exist
LINE 1: select cart_id from old
^
QUERY: select cart_id from old
CONTEXT: PL/pgSQL function generated_id_guard() line 7 at EXECUTE
Aha, «relation "old" does not exist»...
Well, here's the last resort:
drop table cart;
create table cart (
cart_id int generated by default as identity (start with 100) primary key,
at timestamp with time zone
);
insert into cart (cart_id) values (0);
create or replace function surrogate_id_guard() returns trigger
language plpgsql immutable parallel safe cost 1 as $body$
begin
raise exception
'trigger %: updating is prohibited for %',
tg_name, tg_argv[0] using
errcode = 'restrict_violation';
return null;
end;
$body$;
create or replace trigger cart_id_guard
before update of cart_id on cart for each row
execute function surrogate_id_guard('cart_id');
I just make it trigger on any attempt to update cart_id. Let's check:
> update cart set cart_id = 0 where cart_id = 0;
ERROR: trigger cart_id_guard: updating is prohibited for cart_id
CONTEXT: PL/pgSQL function surrogate_id_guard() line 3 at RAISE
> update cart set at = now() where cart_id = 0;
UPDATE 1
Well, finally I answered my original question at this point. But another question is still arisen: How to apply the same algorithm encoded in a function to columns given in args to that function?

The very first attempt in my previous research was to revoke privileges. As Laurenz Albe pointed in his comment I had to revoke the privilege to update the whole table instead of revoking the privilege to update a certain column. Here's the code:
# \c danissimo danissimo
You are now connected to database "danissimo" as user "danissimo".
create table cart (
cart_id int generated by default as identity (start with 100) primary key,
at timestamp with time zone default now()
);
insert into cart default values;
revoke update on cart from danissimo;
Can I update the table now?
> update cart set at = at - interval '1 day';
ERROR: permission denied for table cart
Okay, let's grant the privilege to update columns other than cart_id:
> grant update (at) on cart to danissimo;
> update cart set at = at - interval '1 day';
UPDATE 1
So far, so good. Now time ticks and eventually danissimo adds another column item_ids:
alter table cart add column item_ids int[];
Can danissimo update the new column now? Keep in mind that the privilege to update the whole table was revoked from him and the privilege to update the new column was not granted:
> update cart set item_ids = array[1, 3, 7 ,5];
ERROR: permission denied for table cart
And if I grant him the privilege?
> grant update (item_ids) on cart to danissimo;
> update cart set item_ids = array[1, 3, 7 ,5];
UPDATE 1
What does that all mean? I considered two approaches. One is to prohibit updates of a column once a value is given to the column. Another is to play with privileges. In our projects it's usual that we add new columns while the projects evolve. If I stick with privileges I have to grant the privilege to update the new column each time I add a new one. On the other hand, if I protect some columns with a trigger I just add new columns and bother no more.
CONCLUSION: Use triggers as shown above 👆🏼.

Related

Postgres Count records inserted/ updated

I'm trying to keey track of a clients database with which we sync. I need to record records_added (INSERTs) and records_updated (UPDATEs) to our table.
I'm using an UPSERT to handle the sync, and a trigger to update a table keeping track of insert/updates.
The issue is counting records that have are updated. I have 40+ columns to check, do I have to put all these in my check logic? Is there a more elegant way?
Section of code in question:
select
case
when old.uuid = new.uuid
and (
old.another_field != new.another_field,
old.and_another_field != new.and_another_field,
-- many more columns here << This is particularly painful
) then 1
else 0
end into update_count;
Reproducible example:
-- create tables
CREATE TABLE IF NOT EXISTS example (uuid serial primary key, another_field int, and_another_field int);
CREATE TABLE IF NOT EXISTS tracker_table (
records_added integer DEFAULT 0,
records_updated integer DEFAULT 0,
created_at date unique
);
-- create function
CREATE OR REPLACE FUNCTION update_records_inserted () RETURNS TRIGGER AS $body$
DECLARE update_count INT;
DECLARE insert_count INT;
BEGIN
-- ---------------- START OF BLOCK IN QUESTION -----------------
select
case
when old.uuid = new.uuid
and (
old.another_field != new.another_field
-- many more columns here
) then 1
else 0
end into update_count;
-- ------------------ END OF BLOCK IN QUESTION ------------------
-- count INSERTs
select
case
when old.uuid is null
and new.uuid is not null then 1
else 0
end into insert_count;
-- --log the counts
-- raise notice 'update %', update_count;
-- raise notice 'insert %', insert_count;
-- insert or update count to tracker table
insert into
tracker_table(
created_at,
records_added,
records_updated
)
VALUES
(CURRENT_DATE, insert_count, update_count) ON CONFLICT (created_at) DO
UPDATE
SET
records_added = tracker_table.records_added + insert_count,
records_updated = tracker_table.records_updated + update_count;
RETURN NEW;
END;
$body$ LANGUAGE plpgsql;
-- Trigger
DROP TRIGGER IF EXISTS example_trigger ON example;
CREATE TRIGGER example_trigger
AFTER
INSERT
OR
UPDATE
ON example FOR EACH ROW EXECUTE PROCEDURE update_records_inserted ();
-- A query to insert, then update when number of uses > 1
insert into example(whatever) values (2, 3) ON CONFLICT(uuid) DO UPDATE SET another_field=excluded.another_field+1;

How to implement a constraint in postgresql

In a Details table (Product ID, Receipt ID) only products with a stock greater than 0 can be added.
Warehouse (Product ID, Stock). How can I implement this constraint in Postgresql?
-- CREATE THE FUNCTION
CREATE FUNCTION trg_product_stock_check()
RETURNS trigger AS
$func$
BEGIN
if exists (select * from warehouse w where w.product_id = new.product_id and stock <= 0) then
raise NOTICE 'Product must have a stock of greater than 0';
return null;
end if;
return new;
END
$func$ LANGUAGE plpgsql;
-- CREATE THE TRIGGER
CREATE TRIGGER product_stock_check
BEFORE INSERT ON "orders"
FOR EACH ROW EXECUTE PROCEDURE trg_product_stock_check();
The solution is to create a trigger function that's set up on insert of the order's table. First create your function. You want to check your warehouse table for the product's stock quantity in the function. If the product's stock quantity is less than or equal to 0, return null with a log message. Otherwise, return the newly created row (eg: the new keyword in the function is the row being inserted). After you create the function, you can set the trigger up to run on the orders table before insert.
You can read more about creating a trigger/function here and here.

postgresql trigger to make name unique

I'm using postgres 9.4; I have a table with a unique index. I would like to mutate the name by adding a suffix to ensure the name is unique.
I have created a "before" trigger which computes a suffix. It works well in autocommit mode. However, if two items with the same name are inserted in the same transaction, they both get the same unique suffix.
What is the best way to accomplish my task? Is there a way to handle it with a trigger, or should I ... hmm... wrap the insert or update in a savepoint and then handle the error?
UPDATE (re comment from #Haleemur Ali ):
I don't think my question depends on the details. The salient point is that I query the subset of the collection over which I want to enforce uniqueness, and
choose a new name... however, it would seem that when the queries are run on two objects identically named in the same transaction, one doesn't see the others' modification to the new value.
But ... just in case... my trigger contains ("type" is fixed parameter to the trigger function):
select find_unique(coalesce(new.name, capitalize(type)),
'vis_operation', 'name', format(
'sheet_id = %s', new.sheet_id )) into new.name;
Where "find_unique" contains:
create or replace function find_unique(
stem text, table_name text, column_name text, where_expr text = null)
returns text language plpgsql as $$
declare
table_nt text = quote_ident(table_name);
column_nt text = quote_ident(column_name);
bstem text = replace(btrim(stem),'''', '''''');
find_re text = quote_literal(format('^%s(( \d+$)|$)', bstem));
xtct_re text = quote_literal(format('^(%s ?)', bstem));
where_ext text = case when where_expr is null then '' else 'and ' || where_expr end;
query_exists text = format(
$Q$ select 1 from %1$s where btrim(%2$s) = %3$s %4$s $Q$,
table_nt, column_nt, quote_literal(bstem), where_ext );
query_max text = format($q$
select max(coalesce(nullif(regexp_replace(%1$s, %4$s, '', 'i'), ''), '0')::int)
from %2$s where %1$s ~* %3$s %5$s
$q$,
column_nt, table_nt, find_re, xtct_re, where_ext );
last int;
i int;
begin
-- if no exact match, use exact
execute query_exists;
get diagnostics i = row_count;
if i = 0 then
return coalesce(bstem, capitalize(right(table_nt,4)));
end if;
-- find stem w/ number, use max plus one.
execute query_max into last;
if last is null then
return coalesce(bstem, capitalize(right(table_nt,4)));
end if;
return format('%s %s', bstem, last + 1);
end;
$$;
A BEFORE trigger sees rows modified by the statement that is currently running. So this should work. See demo below.
However, your design will not work in the presence of concurrency. You have to LOCK TABLE ... IN EXCLUSIVE MODE the table you're updating, otherwise concurrent transactions could get the same suffix. Or, with a UNIQUE constraint present, all but one will error out.
Personally I suggest:
Create a side table with the base names and a counter
When you create an entry, lock the side table in EXCLUSIVE mode. This will serialize all sessions that create entries, which is necessary so that you can:
UPDATE side_table SET counter = counter + 1 WHERE name = $1 RETURNING counter to get the next free ID. If you get zero rows, then instead:
Create a new entry in the side table if the base name being created and the counter set to zero.
Demo showing that BEFORE triggers can see rows inserted in the same statement, though not the row that fired the trigger:
craig=> CREATE TABLE demo(id integer);
CREATE TABLE
craig=> \e
CREATE FUNCTION
craig=> CREATE OR REPLACE FUNCTION demo_tg() RETURNS trigger LANGUAGE plpgsql AS $$
DECLARE
row record;
BEGIN
FOR row IN SELECT * FROM demo
LOOP
RAISE NOTICE 'Row is %',row;
END LOOP;
IF tg_op = 'DELETE' THEN
RETURN OLD;
ELSE
RETURN NEW;
END IF;
END;
$$;
CREATE FUNCTION
craig=> CREATE TRIGGER demo_tg BEFORE INSERT OR UPDATE OR DELETE ON demo FOR EACH ROW EXECUTE PROCEDURE demo_tg();
CREATE TRIGGER
craig=> INSERT INTO demo(id) VALUES (1),(2);
NOTICE: Row is (1)
INSERT 0 2
craig=> INSERT INTO demo(id) VALUES (3),(4);
NOTICE: Row is (1)
NOTICE: Row is (2)
NOTICE: Row is (1)
NOTICE: Row is (2)
NOTICE: Row is (3)
INSERT 0 2
craig=> UPDATE demo SET id = id + 100;
NOTICE: Row is (1)
NOTICE: Row is (2)
NOTICE: Row is (3)
NOTICE: Row is (4)
NOTICE: Row is (2)
NOTICE: Row is (3)
NOTICE: Row is (4)
NOTICE: Row is (101)
NOTICE: Row is (3)
NOTICE: Row is (4)
NOTICE: Row is (101)
NOTICE: Row is (102)
NOTICE: Row is (4)
NOTICE: Row is (101)
NOTICE: Row is (102)
NOTICE: Row is (103)
UPDATE 4
craig=>

Trigger to ensure unique values in column

I have to write a trigger to ensure unique entries in column account of table accounts:
create table accounts (id serial, account int4 default 0);
I tried to write my trigger like this:
create function x_6 () returns trigger as '
begin
IF row(new.account) is distinct from row(OLD.account) THEN
return NEW;
ELSE
raise notice '' Entries are not unique! '';
END IF;
end;
'
language 'plpgsql';
Or:
create function x_6 () returns trigger as '
begin
IF (new.account <> OLD.account) THEN
return NEW;
ELSE
raise notice '' Entries are not unique ! '';
END IF;
end;
'
language 'plpgsql';
And then
create trigger x_6t before insert on accounts for each row execute procedure x_6();
When I try to insert something:
insert into accounts(account) values (20);
I get an error in either case:
ERROR: record "old" is not assigned yet
DETAIL: The tuple structure of a not-yet-assigned record is indeterminate.
CONTEXT: PL/pgSQL function "x_6" line 3 at if
How can I fix it?
This is absolutely wrong way. You should not to use triggers for this purpose, you should to use unique indexes.
CREATE TABLE foo(a int PRIMARY KEY, b int);
-- column b has only unique values
CREATE UNIQUE INDEX ON foo(b);
Your code has more than one issue:
bad identifiers - konto instead account
it is table trigger - you has no any access to data there - PostgreSQL triggers are different than MSSQL
If you use row trigger, where there is possible access to record data, then OLD has different meaning than you expect. It is value of record before change - and this value is defined only for UPDATE or DELETE operations - and it is undefined for INSERT, because there previous value of record doesn't exist.

postgres count from table efficient way

In my application we are using postgresql,now it has one million records in summary table.
When I run the following query it takes 80,927 ms
SELECT COUNT(*) AS count
FROM summary_views
GROUP BY question_id,category_type_id
Is there any efficient way to do this?
COUNT(*) in PostgreSQL tends to be slow. It's a feature of MVCC. One of the workarounds of the problem is a row counting trigger with a helper table:
create table table_count(
table_count_id text primary key,
rows int default 0
);
CREATE OR REPLACE FUNCTION table_count_update()
RETURNS trigger AS
$BODY$
begin
if tg_op = 'INSERT' then
update table_count set rows = rows + 1
where table_count_id = TG_TABLE_NAME;
elsif tg_op = 'DELETE' then
update table_count set rows = rows - 1
where table_count_id = TG_TABLE_NAME;
end if;
return null;
end;
$BODY$
LANGUAGE 'plpgsql' VOLATILE;
Next step is to add proper trigger declaration for each table you'd like to use it with. For example for table tab_name:
begin;
insert into table_count values
('tab_name',(select count(*) from tab_name));
create trigger tab_name_table_count after insert or delete
on tab_name for each row execute procedure table_count_update();
commit;
It is important to run in a transaction block to keep actual count and helper table in sync in case of delete or insert between initial count and trigger creation. Transaction guarantees this. From now on to get current count instantly, just invoke:
select rows from table_count where table_count_id = 'tab_name';
Edit: In case of your group by clause, you'll need more sophisticated trigger function and count table.