Re: BUG #19037: Planner fails on estimating array length with "no relation entry" error

Richard Guo <guofenglinux@gmail.com>

From: Richard Guo <guofenglinux@gmail.com>

To: Tom Lane <tgl@sss.pgh.pa.us>

Cc: exclusion@gmail.com, pgsql-bugs@lists.postgresql.org

Date: 2025-09-03T08:08:28Z

Lists: pgsql-bugs

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

Fix planner error when estimating SubPlan cost
- 8c02d92c6475 13.23 landed
- 160ef51c8dba 14.20 landed
- 53e35fb560ab 15.15 landed
- 79ade5873232 16.11 landed
- f34202f51867 17.7 landed
- ab4a35b4ea9d 18.0 landed
- aba8f61c3091 19 (unreleased) landed

On Wed, Sep 3, 2025 at 10:19 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Richard Guo <guofenglinux@gmail.com> writes:
> > One thing I'm not quite sure about is whether we should backpatch this
> > fix to pre-v17 branches.  Prior to v17, estimate_array_length() wasn't
> > taught to use statistics, so this error isn't reproducible there.
> > OTOH, passing a root without a valid simple_rel_array to
> > cost_qual_eval() still seems potentially unsafe.  What do you think?

> Yeah, "is there any other instance of this problem?" is the $64
> question here.  I was initially thinking v17 is sufficient, but
> the possibility that some extension might be vulnerable makes
> me lean to back-patching further.  Your call ...

I've decided to backpatch this fix to pre-v17 branches: using a root
that lacks a valid simple_rel_array in a cost estimation function
seems like a pitfall waiting to happen.  I've included an explanation
of this backpatch decision in the commit message and have pushed the
fix (the test case is not included in pre-v17 branches though).

- Richard