Re: CREATE TABLE creates a composite type corresponding to the table row, which is and is not there - Mailing list pgsql-hackers

From Erik Wienhold
Subject Re: CREATE TABLE creates a composite type corresponding to the table row, which is and is not there
Date
Msg-id 73jtqglviqor5ozqrthjmau7nj4xi7z4obzaqflf7cvliaf4em@4obr7wqlqkye
Whole thread Raw
In response to Re: CREATE TABLE creates a composite type corresponding to the table row, which is and is not there  ("David G. Johnston" <david.g.johnston@gmail.com>)
Responses Re: CREATE TABLE creates a composite type corresponding to the table row, which is and is not there
List pgsql-hackers
On 2024-03-29 02:42 +0100, David G. Johnston wrote:
> For consideration for the doc portion.  The existing wording is too
> imprecise for my liking and just tacking on "expects...create type" is
> jarring.
> 
> """
> Creates a typed table, which takes it structure from an existing (name
> optionally schema-qualified) stand-alone composite type i.e., one created
> using CREATE TYPE) though it still produces a new composite type as well.
> The table will have a dependency to the referenced type such cascaded alter
> and drop actions on the type will propagate to the table.
> 
> A typed table always has the same column names and data types as the type
> it is derived from, and you cannot specify additional columns.  But the
> CREATE TABLE command can add defaults and constraints to the table, as well
> as specify storage parameters.
> """

Thanks, that sounds better.  I incorporated that with some minor edits
in the attached v3.

> We do use the term "stand-alone composite" in create type so I'm inclined
> to use it instead of "composite created with CREATE TYPE"; especially in
> the error messages; I'm a bit more willing to add the cross-reference to
> create type in the user docs.

Okay, changed in v3 as well.  I used "created with CREATE TYPE" in the
error message because I thought it's clearer to the user.  But I see no
reason for not using "stand-alone" here as well if it's the established
term.

-- 
Erik

Attachment

pgsql-hackers by date:

Previous
From: Thomas Munro
Date:
Subject: Re: AIX support
Next
From: Tom Lane
Date:
Subject: Re: AIX support