Date: Wed, 18 May 2022 07:30:05 -0500
From: Justin Pryzby <pryzby@telsasoft.com>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Peter Eisentraut <peter.eisentraut@enterprisedb.com>,
	Amit Langote <amitlangote09@gmail.com>,
	Japin Li <japinli@hotmail.com>, Zhihong Yu <zyu@yugabyte.com>,
	Simon Riggs <simon.riggs@enterprisedb.com>,
	pgsql-hackers@lists.postgresql.org,
	Tomas Vondra <tomas.vondra@enterprisedb.com>,
	Daniel Westermann <dwe@dbi-services.com>,
	Erik Rijkers <er@xs4all.nl>,
	Jaime Casanova <jcasanov@systemguards.com.ec>,
	Andres Freund <andres@anarazel.de>
Subject: Re: support for MERGE
Message-ID: <20220518123005.GA9638@telsasoft.com>
References: <202203281100.qa7wqdu2ooyy@alvherre.pgsql>
 <202203281506.se4wtcvtvdq3@alvherre.pgsql>
 <20220511163350.GL19626@telsasoft.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220511163350.GL19626@telsasoft.com>
User-Agent: Mutt/1.9.4 (2018-02-28)
Archived-At: 
 <https://www.postgresql.org/message-id/20220518123005.GA9638%40telsasoft.com>
Precedence: bulk

On Wed, May 11, 2022 at 11:33:50AM -0500, Justin Pryzby wrote:
> I suggest to reference the mvcc docs from the merge docs, or to make the merge
> docs themselves include the referenced information.

No comments here ?

> Also, EXPLAIN output currently looks like this:
> 
> | Merge on ex_mtarget t (actual rows=0 loops=1)
> |   Tuples Inserted: 0
> |   Tuples Updated: 50
> |   Tuples Deleted: 0
> |   Tuples Skipped: 0
> 
> Should the "zero" rows be elided from the text output ?
> And/or, should it use a more compact output format ?
> 
> There are two output formats already in use, so the options would look like
> this:
> 
>    Tuples: Inserted: 1  Updated: 2  Deleted: 3  Skipped: 4
> or
>    Tuples: inserted=1 updated=2 deleted=3 skipped=4
> 
> Note double spaces and capitals.
> That's separate from the question about eliding zeros.

Actually, the existing uses suggest that these *aren't* separate.

The cases where 0 output is elided (like Heap Blocks and Buffers) uses "=" and
not ":".  The cases using ":" always show all fields (Sort Method, Buckets,
Hits, Batches).  I'm not sure which is preferable for MERGE, but here's one
way.

diff --git a/src/backend/commands/explain.c b/src/backend/commands/explain.c
index b902ef30c87..491105263ff 100644
--- a/src/backend/commands/explain.c
+++ b/src/backend/commands/explain.c
@@ -4119,10 +4119,20 @@ show_modifytable_info(ModifyTableState *mtstate, List *ancestors,
 			skipped_path = total - insert_path - update_path - delete_path;
 			Assert(skipped_path >= 0);
 
-			ExplainPropertyFloat("Tuples Inserted", NULL, insert_path, 0, es);
-			ExplainPropertyFloat("Tuples Updated", NULL, update_path, 0, es);
-			ExplainPropertyFloat("Tuples Deleted", NULL, delete_path, 0, es);
-			ExplainPropertyFloat("Tuples Skipped", NULL, skipped_path, 0, es);
+			if (es->format == EXPLAIN_FORMAT_TEXT)
+			{
+				ExplainIndentText(es);
+				appendStringInfo(es->str,
+						"Tuples Inserted: %.0f  Updated: %.0f  Deleted: %.0f  Skipped: %.0f\n",
+						insert_path, update_path, delete_path, skipped_path);
+			}
+			else
+			{
+				ExplainPropertyFloat("Tuples Inserted", NULL, insert_path, 0, es);
+				ExplainPropertyFloat("Tuples Updated", NULL, update_path, 0, es);
+				ExplainPropertyFloat("Tuples Deleted", NULL, delete_path, 0, es);
+				ExplainPropertyFloat("Tuples Skipped", NULL, skipped_path, 0, es);
+			}
 		}
 	}