Message-ID: <gh-pgjdbc-pgjdbc-3568-c2772437766@github.com>
From: "sehrope (@sehrope)" <noreply+sehrope@github.com>
To: "pgjdbc/pgjdbc" <noreply+pgjdbc-pgjdbc@github.com>
Date: Wed, 02 Apr 2025 12:40:21 +0000
Subject: Re: [pgjdbc/pgjdbc] PR #3568: use pgType instead of internal name to preserve the schema
In-Reply-To: <gh-pgjdbc-pgjdbc-3568@github.com>
References: <gh-pgjdbc-pgjdbc-3568@github.com>
List-Id: <gh-pgjdbc-pgjdbc.github.com>
X-GitHub-Author-Login: sehrope
X-GitHub-Comment-Id: 2772437766
X-GitHub-Comment-Type: issue_comment
X-GitHub-Issue: 3568
X-GitHub-Repo: pgjdbc/pgjdbc
X-GitHub-Type: comment
X-GitHub-Url: https://github.com/pgjdbc/pgjdbc/pull/3568#issuecomment-2772437766
Content-Type: text/plain; charset=utf-8

I think you left out the SQL part of that comment :D

The more we think about these metadata functions, the more weird and annoying edge cases we're going to find. The only consistent answers I can think of are:

* Always query everything at point of use
* Always cache / infer everything using initial search path etc

The first one sucks as it's likely unneeded 99.99% of the time. The second one will be wrong for any changes to search path etc  (but at least it's consistently wrong...). Maybe anybody that does that stuff should really be querying the catalogs in their application code as they're already doing non-JDBC specific actions right?