Re: Question about semantics of $ variables in json explain plans in 13 - Mailing list pgsql-performance

From Tom Lane
Subject Re: Question about semantics of $ variables in json explain plans in 13
Date
Msg-id 1568643.1702083831@sss.pgh.pa.us
Whole thread Raw
In response to Question about semantics of $ variables in json explain plans in 13  (Jerry Brenner <jbrenner@guidewire.com>)
List pgsql-performance
Jerry Brenner <jbrenner@guidewire.com> writes:
> Is there any documentation on the semantics of $ variables in json explain
> plans for both InitPlans and SubPlans in 13?

I don't think there's anything much in the user-facing docs, which is
somewhat unfortunate because it's confusing: the notation is overloaded.
$N could be a parameter supplied from outside the query (as in your $1,
$2 and $3 in the source text), but it could also be a parameter supplied
from an outer query level to a subplan, or it could be the result value
of an InitPlan.  The numbering of outside-the-query parameters is
disjoint from that of the other kind.

>    - It looks like $0 represents the value from the outer query block when
>    the correlated subquery is evaluated
>    - It looks like $1 represents the result of the subquery evaluation

Yeah, I think you're right here.  $0 evidently corresponds to
qRoot.contactID from the outer plan, and the plan label itself
shows that $1 carries the sub-select's value back out.  This $1
is unrelated to the $1 you wrote in the query text.  (It looks
like this is a custom plan in which "67" was explicitly substituted
for your $3.  Presumably $1 and $2 were replaced as well; we don't
do half-custom plans.)

            regards, tom lane



pgsql-performance by date:

Previous
From: Tom Lane
Date:
Subject: Re: 2 json explain plans for the same query/plan - why does one have constants while the other has parameter markers?
Next
From: Jerry Brenner
Date:
Subject: Re: 2 json explain plans for the same query/plan - why does one have constants while the other has parameter markers?