Re: New EXPLAIN option: ALL - Mailing list pgsql-hackers

From Andres Freund
Subject Re: New EXPLAIN option: ALL
Date
Msg-id 20190507164430.a7tig676hjzdnyyd@alap3.anarazel.de
Whole thread Raw
In response to Re: New EXPLAIN option: ALL  (David Fetter <david@fetter.org>)
Responses Re: New EXPLAIN option: ALL  (David Fetter <david@fetter.org>)
List pgsql-hackers
Hi,

On 2019-05-07 18:34:11 +0200, David Fetter wrote:
> On Tue, May 07, 2019 at 08:41:47AM -0700, Andres Freund wrote:
> > On 2019-05-07 09:30:47 +0200, David Fetter wrote:
> > > It can get a little tedious turning on (or off) all the boolean
> > > options to EXPLAIN, so please find attached a shortcut.
> > 
> > I'm not convinced this is a good idea - it seems likely that we'll
> > add conflicting options at some point, and then this will just be a
> > pain in the neck.
> 
> I already left out FORMAT for a similar reason, namely that it's not a
> boolean, so it's not part of flipping on (or off) all the switches.

Which is already somewhat hard to explain.

Imagine if we had CPU_PROFILE = on (which'd be *extremely*
useful). Would you want that to be switched on automatically?  How about
RECORD_IO_TRACE?  How about DISTINCT_BUFFERS which'd be like BUFFERS
except that we'd track how many different buffers are accessed using HLL
or such? Would also be extremely useful.


> Are you seeing a point in the future where there'd be both mutually
> exclusive boolean options and no principled reason to choose among
> them?  If so, what might it look like?

Yes. CPU_PROFILE_PERF, CPU_PROFILE_VTUNE. And lots more.


Greetings,

Andres Freund



pgsql-hackers by date:

Previous
From: David Fetter
Date:
Subject: Re: New EXPLAIN option: ALL
Next
From: David Fetter
Date:
Subject: Re: [PATCH v1] Add a way to supply stdin to TAP tests