Rails migrations for multiple databases and postgresql relations awareness - postgresql

Assuming all databases of a rails 7.0.3 application are on the same server (be they replicas or not)
Where primary database (app_development) has a table shop
a secondary database (appusers_development) has a table role
there is a third database (applocal_development) that will need to
connect as well.
If a migration is created in the secondary database, where
t.references :shop, index: true, null: false, foreign_key: true
t.references :role, index: true, null: false, foreign_key: true
this migration will not pass, as postgresql has no knowledge of the connection to the other database: ActiveRecord::StatementInvalid: PG::UndefinedTable: ERROR: relation "shops" does not exist
Different users exist for main database deploy_root and replicas deploy_readonly.
It does not appear that this is set out in rails functionalities, thus how can it be accomplished with postgresql

There's no way to have a foreign key constraint (or any constraint) between databases (not in postgresql, and not in any engine AFAIK). The whole idea of a relational database is that the relations are contained within a (single) database. Anything outside of that (arguably other than something like a read replica) needs to be handled outside of that.

Related

Question. Access view from Oracle DB in PostgreSQL and Insert into table in Oracle DB from PostgreSQL

For a long time I have been working only with Oracle Databases and I haven't had much contact with PostgreSQL.
So now, I have a few questions for people who are closer to Postgres.
Is it possible to create a connection from Postgres to Oracle (oracle_fdw?) and perform selects on views in a different schema than the one you connected to?
Is it possible to create a connection from Postgres to Oracle (oracle_fdw?) and perform inserts on tables in the same schema as the one you connected to?
Ad 1:
Yes, certainly. Just define the foreign table as
CREATE FOREIGN TABLE view_1_r (...) SERVER ...
OPTIONS (table 'VIEW_1', schema 'USERB');
Ad 2:
Yes, certainly. Just define a foreign table on the Oracle table and insert into it. Note that bulk inserts work, but won't perform well, since there will be a round trip between PostgreSQL and Oracle for each row inserted.
Both questions indicate a general confusion between a) the Oracle user that you use to establish the connection and b) the schema of the table or view that you want to access. These things are independent: The latter is determined by the schema option of the foreign table definition, while the former is determined by the user mapping.

Create Postgres FDW using the postgres super user yet have read only access to the remote schema

I am trying to create an FDW using the postgres super user credentials using the following steps -
My super user is postgres
CREATE SERVER testfdw
FOREIGN DATA WRAPPER postgres_fdw
OPTIONS (host '10.10.10.171', dbname 'testdb', port '5432', fetch_size '50000');
I have created a read only remote user called testdb_read_only_user on testdb and granted read only privileges to all tables and sequences and functions in there
I now create a user mapping for my local db super user postgres
CREATE USER MAPPING
FOR postgres
SERVER testfdw
OPTIONS (user 'testdb_read_only_user', password 'testpasswrd')
Then I run IMPORT FOREIGN SCHEMA public FROM SERVER testfdw INTO test_local_schema;
Now when I try to create a new foreign table into test_local_schema using the postrges super user, I am still able to create one, but the actual behaviour I expect is a permission denied error and my user mapping is based on the read only user I created on the Remote DB
Am I missing something here. Please guide.
The foreign schema test_local_schema is owned by postgres(Superuser) and PostgreSQL treating it as a local schema(on target) gives you option to create a new table, Now when you are creating a new foreign table in test_local_schema postgres will check below constraints
The name of the foreign table must be distinct from the name of any other foreign table, table, sequence, index, view, or materialized view in the same schema.
To create a foreign table, you must have USAGE(read) privilege on
the foreign server, as well as USAGE privilege on all column
types used in the table
CREATE privilege on the target schema.
As your user mapping has read-only user in source database you should be able to fetch data in order to create a foreign table in target database, however whenever you will update and insert new data it will give you error as testdb_read_only_user is not having any update/insert privileges.
https://www.postgresql.org/docs/9.5/sql-createforeigntable.html
https://www.postgresql.org/docs/9.5/sql-importforeignschema.html

Entity Framework Code First Migration and Data Migrations

We are using Entity Framework Code First with Migrations. We have it running without any issues. We are trying to migrate data from an old database to the new. We need to drop the database, create all the tables, insert the data and then add in the primary keys and foreign keys. We want to do this in EF so that we can format the data being migrated. We have been successful in creating the database and then migrating the data over but the primary keys are in the old database are not coming over. I've tried using the ExecuteSqlCommand before a bulkinsert:
ctx.Database.ExecuteSqlCommand("SET IDENTITY_INSERT [dbo].[TableName] ON");
but this does not work because the migration scipts have the idenity set to true:
Id = c.Int(nullable: false, identity: true),
Is there a way to set the identity to false then after the data is inserted into the database, set the identity to true?
You are doing it in the right way (Id is an identity and you enable identity insert). SET IDENTITY_INSERT ON is valid only for the session (the connection) so you need to use only one connection.
Also, you have to seed the database using INSERT statements (and not EF SaveChanges).

Allow users to create tables, but not drop them in PostgreSQL

By default, PostgreSQL allows the owner and any SUPERUSER roles to drop tables. We're trying to create a role that can create tables in a schema, but not drop them once they're created.
More ideally, we'd actually like the created tables to be owned by postgres.
Is there any way to force created tables in a schema to automatically be owned by postgres, or at least prevent them from being dropped by the owner? We are using PostgreSQL 9.4.

Joining Results from Two Separate Databases

Is it possible to JOIN rows from two separate postgres databases?
I am working with system with couple databases in one server and sometimes I really need such a feature.
According to http://wiki.postgresql.org/wiki/FAQ
There is no way to query a database other than the current one.
Because PostgreSQL loads database-specific system catalogs, it is
uncertain how a cross-database query should even behave.
contrib/dblink allows cross-database queries using function calls. Of
course, a client can also make simultaneous connections to different
databases and merge the results on the client side.
EDIT: 3 years later (march 2014), this FAQ entry has been revised and is more helpful:
How do I perform queries using multiple databases?
There is no way to directly query a database other than the current
one. Because PostgreSQL loads database-specific system catalogs, it is
uncertain how a cross-database query should even behave.
The SQL/MED support in PostgreSQL allows a "foreign data wrapper" to
be created, linking tables in a remote database to the local database.
The remote database might be another database on the same PostgreSQL
instance, or a database half way around the world, it doesn't matter.
postgres_fdw is built-in to PostgreSQL 9.3 and includes read/write
support; a read-only version for 9.2 can be compiled and installed as
a contrib module.
contrib/dblink allows cross-database queries using function calls and
is available for much older PostgreSQL versions. Unlike postgres_fdw
it can't "push down" conditions to the remote server, so it'll often
land up fetching a lot more data than you need.
Of course, a client can also make simultaneous connections to
different databases and merge the results on the client side.
Forget about dblink!
Say hello to Postgres_FDW:
To prepare for remote access using postgres_fdw:
Install the postgres_fdw extension using CREATE EXTENSION.
Create a foreign server object, using CREATE SERVER, to represent each remote database you want to connect to. Specify connection
information, except user, and password, as options of the server
object.
Create a user mapping, using CREATE USER MAPPING, for each database user you want to allow to access each foreign server. Specify
the remote user name and password to use as user and password options
of the user mapping.
Create a foreign table, using CREATE FOREIGN TABLE or IMPORT FOREIGN SCHEMA, for each remote table you want to access. The columns
of the foreign table must match the referenced remote table. You can,
however, use table and/or column names different from the remote
table's, if you specify the correct remote names as options of the
foreign table object.
Now you need only SELECT from a foreign table to access the data
stored in its underlying remote table.
It's really useful even on large data.
Yes, it is possible to do this using dblink albeit with significant performance considerations.
The following example will require the current SQL user to have permissions on both databases. If db2 is not located on the same cluster, then you will need to replace dbname=db2 with the full connection string defined in the dblink documentation.
SELECT *
FROM table1 tb1
LEFT JOIN (
SELECT *
FROM dblink('dbname=db2','SELECT id, code FROM table2')
AS tb2(id int, code text);
) AS tb2 ON tb2.column = tb1.column;
If table2 is very large, you could have performance issues because the sub-query loads up the entire table2 before performing the join.
No you can't. You could use dblink to connect from one database to another database, but that won't help if you're looking for JOIN's.
You can't use different SCHEMA's within a single database to store all you data?
Just a few steps and You can reach the goal:
follow this reference step by step
WE HAVE BEEN CONNECTED TO DB2 WITH TABLE TBL2 AND COLUMN COL2
ALSO THERE IS DB1 WITH TBL1 AND COLUMN COL1
*** connecting to second db ie db2
Now just **copy paste the 1-7 processes** (make sure u use correct username and password and ofcourse db name)
1.**CREATE EXTENSION dblink;**
2.**SELECT pg_namespace.nspname, pg_proc.proname
FROM pg_proc, pg_namespace
WHERE pg_proc.pronamespace=pg_namespace.oid
AND pg_proc.proname LIKE '%dblink%';**
3.**SELECT dblink_connect('host=localhost user=postgres password=postgres dbname=db1');**
4.**CREATE FOREIGN DATA WRAPPER postgres VALIDATOR postgresql_fdw_validator;**
5.**CREATE SERVER postgres2 FOREIGN DATA WRAPPER postgres OPTIONS (hostaddr '127.0.0.1', dbname 'db1');**
6.**CREATE USER MAPPING FOR postgres SERVER postgres2 OPTIONS (user 'postgres', password 'postgres');**
7.**SELECT dblink_connect('postgres2');**
---Now, you can SELECT the data of Database_One from Database_Two and even join both db results:
**SELECT * FROM public.dblink
('postgres2','SELECT col1,um_name FROM public.tbl1 ')
AS DATA(um_userid INTEGER),tbl2 where DATA.col1=tbl2.col2;**
You can also Check this :[How to join two tables of different databases together in postgresql [\[working finely in version 9.4\]][1]
You need to use dblink...as araqnid mentioned above, something like this works fine:
select ST.Table_Name, ST.Column_Name, DV.Table_Name, DV.Column_Name, *
from information_schema.Columns ST
full outer join dblink('dbname=otherdatabase','select Table_Name,
Column_Name from information_schema.Columns') DV(Table_Name text,
Column_Name text)
on ST.Table_Name = DV.Table_name
and ST.Column_Name = DV.Column_Name
where ST.Column_Name is null or DV.Column_Name is NULL
You have use dblink extension of postgresql.
Reference take from this Article:
DbLink extension of PostgreSQL which is used to connect one database to another database.
Install DbLink extension.
CREATE EXTENSION dblink;
Verify DbLink:
SELECT pg_namespace.nspname, pg_proc.proname
FROM pg_proc, pg_namespace
WHERE pg_proc.pronamespace=pg_namespace.oid
AND pg_proc.proname LIKE '%dblink%';
I have already prepared full demonstration on this. Please visit my post to learn step by step for executing cross database query in Postgresql.
Cannot be done? Of course we can, without special extensions. In our case, we had to compare two tables from different database servers, e.g. ACC and PROD, hence an even harder case than from most answers. Especially because ACC and PROD are deliberately on different servers to create a barrier, so you will not easily gain enough rights to perform a GRANT USAGE ON FOREIGN SERVER.
The obvious solution is to export both tables, and import both in the same database, e.g. DEV, or your own local db, under appropriate names, e.g. table1_acc and table1_prod, or schemas like acc and prod. Then, you may JOIN those with no special problems.