proposal: PLpgSQL parallel assignemnt

On 2 June 2017 at 15:51, Pavel Stehule <pavel.stehule@gmail.com> wrote:

a, b := fx();

Comments, notes, ideas?

I'd be pretty happy to have

(a, b) = (x, y);
(a, b) = f(x);

which is SQL-esque.

But what, if anything, does Ada do?

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

2017-06-02 10:06 GMT+02:00 Craig Ringer <craig@2ndquadrant.com>:

On 2 June 2017 at 15:51, Pavel Stehule <pavel.stehule@gmail.com> wrote:

a, b := fx();

Comments, notes, ideas?

I'd be pretty happy to have

(a, b) = (x, y);
(a, b) = f(x);

which is SQL-esque.

This is not too far to my proposal - and it is fully adequate alternative.

But what, if anything, does Ada do?

What I know, no, Ada has not this statement - but the design of OUT
parameters in Ada absolutely different than PostgreSQL - so in this case we
cannot to use Ada language as our base :(

Regards

Pavel

2017-06-02 10:15 GMT+02:00 Pavel Stehule <pavel.stehule@gmail.com>:

2017-06-02 10:06 GMT+02:00 Craig Ringer <craig@2ndquadrant.com>:

On 2 June 2017 at 15:51, Pavel Stehule <pavel.stehule@gmail.com> wrote:

a, b := fx();

Comments, notes, ideas?

I'd be pretty happy to have

(a, b) = (x, y);
(a, b) = f(x);

which is SQL-esque.

This is not too far to my proposal - and it is fully adequate alternative.

The ANSI form is related to SET or UPDATE commands - so in this case I see
classic languages style
https://en.wikipedia.org/wiki/Assignment_(computer_science) better. The
assign statement in PLpgSQL is not related to embedded SQL. If we introduce
SQL syntax and SET commands for schema variables then ( ) syntax is
perfect, but for := PLpgSQL I am not sure

It is maybe strange, but

SET (a,b) = (SELECT a,b FROM foo)

a, b := fx()

are sentences from two independent worlds and different syntax can be
correct (depends how much we would to integrate procedural and SQL worlds
.. 100% T-SQL, 80% SQL/PSM, ..20% PLpgSQL or 5%PL/SQL)

Regards

Pavel

2017-06-02 19:05 GMT+02:00 Pavel Stehule <pavel.stehule@gmail.com>:

2017-06-02 10:15 GMT+02:00 Pavel Stehule <pavel.stehule@gmail.com>:

2017-06-02 10:06 GMT+02:00 Craig Ringer <craig@2ndquadrant.com>:

On 2 June 2017 at 15:51, Pavel Stehule <pavel.stehule@gmail.com> wrote:

a, b := fx();

Comments, notes, ideas?

I'd be pretty happy to have

(a, b) = (x, y);
(a, b) = f(x);

which is SQL-esque.

This is not too far to my proposal - and it is fully adequate
alternative.

The ANSI form is related to SET or UPDATE commands - so in this case I see
classic languages style https://en.wikipedia.org/wiki/
Assignment_(computer_science) better. The assign statement in PLpgSQL is
not related to embedded SQL. If we introduce SQL syntax and SET commands
for schema variables then ( ) syntax is perfect, but for := PLpgSQL I am
not sure

It is maybe strange, but

SET (a,b) = (SELECT a,b FROM foo)

a, b := fx()

are sentences from two independent worlds and different syntax can be
correct (depends how much we would to integrate procedural and SQL worlds
.. 100% T-SQL, 80% SQL/PSM, ..20% PLpgSQL or 5%PL/SQL)

More thoughts:

1. syntax (a,b) := f() ... can mean - assign record to temporary composite
(a,b)
2. syntax a,b := f() ... can mean - unpack result composite and assign to
a, b fields

so both syntaxes has sense although we don't introduce relation to SQL - on
this way

a,b := 10, 20 -- ok .. attach a=c1, b=c2
a,b := (10,20) -- ok .. attach a = r.c1, b = r.c2
(a,b) := (10,20) -- ok attach ct = rt
(a,b) := 10,20 -- ok attach ct = row(c1, c2)

@1 syntax says "create composite target", @2 syntax says "unpack result".
Both should to work. Personally I prefer @1 .. due less parenthesis

Regards

Pavel