Re: Changing a schema's name with function1 calling function2 - Mailing list pgsql-general

From Adrian Klaver
Subject Re: Changing a schema's name with function1 calling function2
Date
Msg-id f09b9fb9-98b9-4d39-a47b-c32ca3ff8782@aklaver.com
Whole thread Raw
In response to Re: Re: Changing a schema's name with function1 calling function2  ("Wilma Wantren" <wilma.wantren@eclipso.de>)
Responses Re: Re: Changing a schema's name with function1 calling function2
List pgsql-general
On 12/22/23 15:09, Wilma Wantren wrote:
> Thank you for your response!
> No, this does not help me. I will try to explain better what I mean. Our customers use our program with a Postgres
database,user and schema, which have the same name. The customer chooses the name. Let's say one customer calls the
database,user and schema 'my_things'.
 
> I want to create scripts for the customers with which they can export the schema and import it into another database
witha new schema name, let's say one customer wants to import it as 'my_things1' (this schema belongs to database
'my_things1').
> The export script calls pg_dump to export schema 'my_things'.
> The import script calls pg_restore to restore schema 'my_things' in database 'my_things1' and then calls psql to
changethe schema name to 'my_things1'.
 
> Now there is function1 which its search_path set to 'my_things'. Because the search_path is still set to 'my_things'
afterrenaming the schema the script must now call psql to change the function's search_path to 'my_things1'.
 
> This is not just one line more in the import script.
> It is a fact that I must know - if I did not know about function1's search_path then there would be an error in the
schemaafter renaming the schema.
 
> And imagine if one day a colleague of mine implements a new function which needs a search_path but the colleague
forgetsto adjust the import script then again there is an error in the schema after renaming the schema.
 
> Therefore it would be great if there was a variable which I could set in a search_path (like the variable "$user")
whichdenotes the function's schema and which is only evaluated when the function is executed, i.e. the variable would
bethe value of the function's search_path in the function's meta data. This variable, e.g. "$function_schema" would
stilldenote the correct schema after renaming the schema.
 
> 

There is no such variable. What you have available is what I showed before:

https://www.postgresql.org/docs/current/sql-alterfunction.html

"
configuration_parameter
value

      Add or change the assignment to be made to a configuration
parameter when the function is called. If value is DEFAULT or,
equivalently, RESET is used, the function-local setting is removed, so
that the function executes with the value present in its environment.
Use RESET ALL to clear all function-local settings. SET FROM CURRENT
saves the value of the parameter that is current when ALTER FUNCTION is

executed as the value to be applied when the function is entered.

      See SET and Chapter 20 for more information about allowed parameter

names and values.
"

The alternative is to create a migration process using scripts with 
variable substitution outside the database. I do something similar using 
Sqitch(https://sqitch.org/) and its template system:

https://sqitch.org/docs/manual/sqitch-add/

-- 
Adrian Klaver
adrian.klaver@aklaver.com




pgsql-general by date:

Previous
From: "Wilma Wantren"
Date:
Subject: Re: Re: Changing a schema's name with function1 calling function2
Next
From: Phillip Diffley
Date:
Subject: Re: How to generate random bigint