Add red-black tree missing comparison searches

Please add this to the commitfest at
https://commitfest.postgresql.org/38/ so it doesn't get missed. The
commitfest starts imminently so best add it today.

Yes, I've already added it here: https://commitfest.postgresql.org/38/3742/

Thanks!

On Thu, 30 Jun 2022 at 12:09, Greg Stark <stark@mit.edu> wrote:

Hi, Steve!

Thank you for working on this.

On Thu, Jun 30, 2022 at 7:51 PM Steve Chavez <steve@supabase.io> wrote:

Currently the red-black tree implementation only has an equality search. Other extensions might need other comparison searches, like less-or-equal or greater-or-equal. For example OrioleDB defines a greater-or-equal search on its postgres fork:

https://github.com/orioledb/postgres/blob/4c18ae94c20e3e95c374b9947f1ace7d1d6497a1/src/backend/lib/rbtree.c#L164-L186

So I thought this might be valuable to have in core. I've added less-or-equal and greater-or-equal searches functions plus tests in the test_rbtree module. I can add the remaining less/great searches if this is deemed worth it.

Looks good. But I think we can support strict inequalities too (e.g.
less and greater without equals). Could you please make it a bool
argument equal_matches?

Also I refactored the sentinel used in the rbtree.c to use C99 designators.

Could you please extract this change as a separate patch.

------
Regards,
Alexander Korotkov

Hey Alexander,

But I think we can support strict inequalities too (e.g.

less and greater without equals). Could you please make it a bool
argument equal_matches?

Sure, an argument is a good idea to keep the code shorter.

Could you please extract this change as a separate patch.

Done!

On Thu, 30 Jun 2022 at 14:34, Alexander Korotkov <aekorotkov@gmail.com>
wrote:

Hi, Steve!

On Sat, Jul 2, 2022 at 10:38 PM Steve Chavez <steve@supabase.io> wrote:

But I think we can support strict inequalities too (e.g.

less and greater without equals). Could you please make it a bool
argument equal_matches?

Sure, an argument is a good idea to keep the code shorter.

Could you please extract this change as a separate patch.

Done!

Thank you!

I did some improvements to the test suite, run pgindent and wrote
commit messages.

I think this is quite straightforward and low-risk patch. I'm going
to push it if no objections.

------
Regards,
Alexander Korotkov

---------- Forwarded message ---------
From: Steve Chavez <steve@supabase.io>
Date: Wed, 6 Jul 2022 at 18:14
Subject: Re: Add red-black tree missing comparison searches
To: Alexander Korotkov <aekorotkov@gmail.com>

Thanks Alexander!

wrt to the new patch. I think the following comment is misleading since
keyDeleted can be true or false:

+ /* switch equal_match to false so we only find greater matches now */
+ node = (IntRBTreeNode *) rbt_find_great(tree, (RBTNode *) &searchNode,
+ keyDeleted);

Maybe it should be the same used for searching lesser keys:

+ /*
+ * Find the next key.  If the current key is deleted, we can pass
+ * equal_match == true and still find the next one.
+ */

On Wed, 6 Jul 2022 at 13:53, Alexander Korotkov <aekorotkov@gmail.com>
wrote:

On Thu, Jul 7, 2022 at 2:16 AM Steve Chavez <steve@supabase.io> wrote:

Thanks Alexander!

wrt to the new patch. I think the following comment is misleading since keyDeleted can be true or false:

+ /* switch equal_match to false so we only find greater matches now */
+ node = (IntRBTreeNode *) rbt_find_great(tree, (RBTNode *) &searchNode,
+ keyDeleted);

Maybe it should be the same used for searching lesser keys:

+ /*
+ * Find the next key.  If the current key is deleted, we can pass
+ * equal_match == true and still find the next one.
+ */

Thank you for catching this.
The revised version of patch is attached!

------
Regards,
Alexander Korotkov

On Thu, Jul 7, 2022 at 1:43 PM Alexander Korotkov <aekorotkov@gmail.com> wrote:

On Thu, Jul 7, 2022 at 2:16 AM Steve Chavez <steve@supabase.io> wrote:

Thanks Alexander!

wrt to the new patch. I think the following comment is misleading since keyDeleted can be true or false:

+ /* switch equal_match to false so we only find greater matches now */
+ node = (IntRBTreeNode *) rbt_find_great(tree, (RBTNode *) &searchNode,
+ keyDeleted);

Maybe it should be the same used for searching lesser keys:

+ /*
+ * Find the next key.  If the current key is deleted, we can pass
+ * equal_match == true and still find the next one.
+ */

Thank you for catching this.
The revised version of patch is attached!

Pushed!

------
Regards,
Alexander Korotkov