Re: Transition relation clarification - Mailing list pgsql-docs

From Dull Bananas
Subject Re: Transition relation clarification
Date
Msg-id CAKmgnAhPLHwg3v4FH3t9PHG-xt0pZJHMYnLPrG=4GWCB3Nu=Zw@mail.gmail.com
Whole thread Raw
In response to Re: Transition relation clarification  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-docs
The documentation is not specific enough about what a transition table is or what it means to create a transition table, so I assumed that it works similarly to `CREATE TEMPORARY TABLE`. The existence of variable-like tables is not documented.

On Mon, Jan 1, 2024 at 10:22 David G. Johnston <david.g.johnston@gmail.com> wrote:
On Mon, Jan 1, 2024 at 9:35 AM Dull Bananas <dull.bananas0@gmail.com> wrote:
The docs should mention that transition relations aren't automatically made available in functions called by the trigger function. This unknown behavior caused a lot of frustration for me.

Seems to fall into the "if it isn't documented it doesn't happen" category of things that we simply choose not to clutter the documentation with.  Called functions don't inherit stuff from the caller.  It is like documenting a variable in the calling function isn't visible to the called function.  Admittedly at least with variables you can pass them via arguments.  The special trigger state is indeed available in the trigger returning function when it is the function directly invoked by the trigger.

David J.

pgsql-docs by date:

Previous
From: "David G. Johnston"
Date:
Subject: Re: Transition relation clarification
Next
From: Tom Lane
Date:
Subject: Re: Postgres compilation instructions do not work on modern debian or ubuntu