pl/pgsql oddity

[Joolz]

Hello everyone,

When writing some serverside code I ran into an oddity that I
managed to boil down to this:

-------------------------------------------------------
create or replace function fubar() returns varchar as '
declare
l integer;
begin
l = 38;
if l < 38 then
return ''< 38'';
elseif l >= 38 then
return ''>= 38'';
else
return ''this is not possible'';
end if;
end;'
language 'plpgsql';
-------------------------------------------------------

But I can't understand or solve it. The function always returns
'this is not possible'. Can someone tell me what I'm overlooking?
Thanks!


---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

[url]http://www.postgresql.org/docs/faqs/FAQ.html[/url]

[Matteo Beccati]

Hi,

> l = 38;

This should be:

l := 38;

otherwise l would remain uninitialized (NULL).


Ciao ciao
--
Matteo Beccati
[url]http://phpadsnew.com[/url]
[url]http://phppgads.com[/url]


---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

[url]http://www.postgresql.org/docs/faqs/FAQ.html[/url]

[Ian Barwick]

On Thu, 16 Dec 2004 10:06:19 +0100 (CET), Joolz
<joolz@arbodienst-limburg.nl> wrote:

> Hello everyone,
>
> When writing some serverside code I ran into an oddity that I
> managed to boil down to this:
>
> -------------------------------------------------------
> create or replace function fubar() returns varchar as '
> declare
> l integer;
> begin
> l = 38;
> if l < 38 then
> return ''< 38'';
> elseif l >= 38 then

Try "elsif" here.
(No, I don't know what the problem with "elseif" is).

Ian Barwick

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

[Richard Huxton]

Joolz wrote:

> Hello everyone,
>
> When writing some serverside code I ran into an oddity that I
> managed to boil down to this:

> elseif l >= 38 then

You want "elsif" - plpgsql isn't a hugely sophisticated language and its
parser is having trouble there. I'm guessing the parser is somehow
putting the "elseif" branch under the initial "then" so it never gets
executed. If you rewrite the function like so:

create or replace function fubar() returns varchar as '
declare
l integer;
begin
l = 34;
if l < 38 then
raise notice ''< 38: %'',l;
elseif l >= 38
then raise notice ''>= 38: %'',l;
else
raise notice ''this is not possible: %'',l;
end if;

return 0;
end;'
language 'plpgsql';

Now, try different values for "l" and you'll see what is happening.
Congratulations - I think you've found a bug. You can report it formally
via the bugs mailing list or [url]http://www.postgresql.org/bugform.html[/url]

--
Richard Huxton
Archonet Ltd

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to [email]majordomo@postgresql.org[/email]

[Joolz]

Ian Barwick zei:

> On Thu, 16 Dec 2004 10:06:19 +0100 (CET), Joolz
> <joolz@arbodienst-limburg.nl> wrote:

>> Hello everyone,
>>
>> When writing some serverside code I ran into an oddity that I
>> managed to boil down to this:
>>
>> -------------------------------------------------------
>> create or replace function fubar() returns varchar as '
>> declare
>> l integer;
>> begin
>> l = 38;
>> if l < 38 then
>> return ''< 38'';
>> elseif l >= 38 then

>
> Try "elsif" here.
> (No, I don't know what the problem with "elseif" is).

Thanks Ian, but I don't understand... I _am_ already using elseif...


---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

[Tomasz Myrta]

> When writing some serverside code I ran into an oddity that I

> managed to boil down to this:
>
> -------------------------------------------------------
> create or replace function fubar() returns varchar as '
> declare
> l integer;
> begin
> l = 38;
> if l < 38 then
> return ''< 38'';
> elseif l >= 38 then
> return ''>= 38'';
> else
> return ''this is not possible'';
> end if;
> end;'
> language 'plpgsql';
> -------------------------------------------------------
>
> But I can't understand or solve it. The function always returns
> 'this is not possible'. Can someone tell me what I'm overlooking?
> Thanks!

Documentation says you should use "elsif" rather than "elseif". In your
case everything between "return <38" and "else" is discarded as
unreachable code. The same sense will have function below:

create or replace function fubar() returns varchar as '
declare
l integer;
begin
l = 38;
if l < 38 then
return ''< 38'';
bleblbebleblebe sfdsdf;
nothing special;
else
return ''this is not possible'';
end if;
end;'
language 'plpgsql';
select fubar();

Regards,
Tomasz Myrta

[Joolz]

Tomasz Myrta zei:

>> When writing some serverside code I ran into an oddity that I
>> managed to boil down to this:
>>
>> -------------------------------------------------------
>> create or replace function fubar() returns varchar as '
>> declare
>> l integer;
>> begin
>> l = 38;
>> if l < 38 then
>> return ''< 38'';
>> elseif l >= 38 then
>> return ''>= 38'';
>> else
>> return ''this is not possible'';
>> end if;
>> end;'
>> language 'plpgsql';
>> -------------------------------------------------------
>>
>> But I can't understand or solve it. The function always returns
>> 'this is not possible'. Can someone tell me what I'm overlooking?
>> Thanks!

>
> Documentation says you should use "elsif" rather than "elseif". In
> your
> case everything between "return <38" and "else" is discarded as
> unreachable code.

That's it, thanks!



---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [email]majordomo@postgresql.org[/email] so that your
message can get through to the mailing list cleanly

[Joolz]

Richard Huxton zei:

Hi Richard,

See the other posting,

elseif l >= 38

Apparently this is parsed as

elseif l >= 38
^ ^
| |
code|
|
comment from here on

It should be "elsif", not "elseif" :-\

Thanks everyone!


---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [email]majordomo@postgresql.org[/email])

[Neil Conway]

Richard Huxton wrote:

> You want "elsif" - plpgsql isn't a hugely sophisticated language and its
> parser is having trouble there. I'm guessing the parser is somehow
> putting the "elseif" branch under the initial "then" so it never gets
> executed.

Indeed; the parser thinks an unrecognized keyword indicates the
beginning of a SQL statement: since the PL/PgSQL parser and the SQL
parser are completely separate, we need to do some guessing about what
constitutes a legal SQL statement. See the more detailed diagnose of the
problem here:

[url]http://archives.postgresql.org/pgsql-bugs/2004-11/msg00297.php[/url]

There's a patch in that thread that provides better PL/PgSQL error
checking (which results in flagging this kind of code as invalid at
compile time). Some form of that patch will be in 8.1, as well as other
nice stuff like warning for unreachable code.

Tom also suggested just adding 'elseif' as an alternative for 'elsif'.
That sounds like it would be worth doing.

> Congratulations - I think you've found a bug. You can report it formally
> via the bugs mailing list

No need, this is a known issue.

-Neil

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [email]majordomo@postgresql.org[/email] so that your
message can get through to the mailing list cleanly

[Richard Huxton]

Matteo Beccati wrote:

> Hi,
>

>> l = 38;

>
>
> This should be:
>
> l := 38;
>
> otherwise l would remain uninitialized (NULL).

Actually, either work. You are right that the docs suggest the second
form though.

--
Richard Huxton
Archonet Ltd

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to [email]majordomo@postgresql.org[/email]

[Tom Lane]

Neil Conway <neilc@samurai.com> writes:

> Tom also suggested just adding 'elseif' as an alternative for 'elsif'.
> That sounds like it would be worth doing.

I think we should go ahead and do that for 8.0. I'm getting tired of
reading reports that stem from this mistake (I think this is the third
one in the past month ...). I can't see any real downside to accepting
both spellings, can you?

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

[url]http://www.postgresql.org/docs/faqs/FAQ.html[/url]

[Geoffrey]

Tom Lane wrote:

> Neil Conway <neilc@samurai.com> writes:
>

>>Tom also suggested just adding 'elseif' as an alternative for 'elsif'.
>>That sounds like it would be worth doing.

>
>
> I think we should go ahead and do that for 8.0. I'm getting tired of
> reading reports that stem from this mistake (I think this is the third
> one in the past month ...). I can't see any real downside to accepting
> both spellings, can you?

Well, you end up with the opposite problem. Someone has a issue and is
focusing on the spelling of elseif (elsif), because the google searches
they've done happen to come up with a different spelling then what
they're using. So, they're looking for a solution in the wrong place.

I don't know of any other language that permits multiple spellings for
the same construct. I'd be concerned with starting such a precedent.

Now I'm sure someone will post back with some valid examples from other
languages. :(

--
Until later, Geoffrey

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

[url]http://archives.postgresql.org[/url]

[Michael Fuhr]

On Thu, Dec 16, 2004 at 12:27:53PM -0500, Geoffrey wrote:

> I don't know of any other language that permits multiple spellings for
> the same construct. I'd be concerned with starting such a precedent.

I'd be in favor of making it a bloody law that every bloody language
use the same bloody spelling. I'm forever forgetting whether a
particular language uses ELSE IF, ELSEIF, ELSIF, or ELIF. Grumble,
grumble, grumble....

--
Michael Fuhr
[url]http://www.fuhr.org/~mfuhr/[/url]

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [email]majordomo@postgresql.org[/email] so that your
message can get through to the mailing list cleanly

[Tom Lane]

Geoffrey <esoteric@3times25.net> writes:

> I don't know of any other language that permits multiple spellings for
> the same construct. I'd be concerned with starting such a precedent.

Well, we have plenty of precedent already at the SQL language level:
ANALYZE vs ANALYSE, NOTNULL vs IS NOT NULL, and so on.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

[url]http://archives.postgresql.org[/url]

[Geoffrey]

Tom Lane wrote:

> Geoffrey <esoteric@3times25.net> writes:
>

>>I don't know of any other language that permits multiple spellings for
>>the same construct. I'd be concerned with starting such a precedent.

>
>
> Well, we have plenty of precedent already at the SQL language level:
> ANALYZE vs ANALYSE, NOTNULL vs IS NOT NULL, and so on.

Like I said, someone would come up with some real world examples. :) I
personally wouldn't lose too much sleep over it.

--
Until later, Geoffrey

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [email]majordomo@postgresql.org[/email] so that your
message can get through to the mailing list cleanly

[Guy Rouillier]

Michael Fuhr wrote:

> On Thu, Dec 16, 2004 at 12:27:53PM -0500, Geoffrey wrote:
>

>> I don't know of any other language that permits multiple spellings
>> for the same construct. I'd be concerned with starting such a
>> precedent.

>
> I'd be in favor of making it a bloody law that every bloody language
> use the same bloody spelling. I'm forever forgetting whether a
> particular language uses ELSE IF, ELSEIF, ELSIF, or ELIF. Grumble,
> grumble, grumble....

As a relative newbie to PostgreSQL (but an old-timer to programming
languages and other DMBSs) I would certainly vote for allowing elseif.
This is my first encounter of "else" without the terminating "e", and
that would not be a natural omission for me.

--
Guy Rouillier


---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

[Frank D. Engel, Jr.]

"elsif" is the spelling used by Ada. I'm getting rather used to it,
myself.

I'm really starting to like Ada. So "elsif" is fine with me.

As far as alternate spellings being accepted within a language, look at
the Transcript language used by Runtime Revolution ([url]www.runrev.com[/url]),
which is a so-called "Xtalk" language (based on HyperTalk, which was
used by HyperCard; and yes, HyperTalk was like this too...)

-- and # are synonyms (comment delimiters)

abbreviated can be abbreviated as abbr or abbrev ;-)

audioClip can be abbreviated as ac

accelKey and acceleratorKey

acceleratorModifiers and accelMods

arm and armed

autoHilite and autoHighlight

recentNames and backList

group, background, bg, and bkgnd


The list goes on and On and ON...


On Dec 16, 2004, at 2:12 PM, Guy Rouillier wrote:

> Michael Fuhr wrote:

>> On Thu, Dec 16, 2004 at 12:27:53PM -0500, Geoffrey wrote:
>>

>>> I don't know of any other language that permits multiple spellings
>>> for the same construct. I'd be concerned with starting such a
>>> precedent.

>>
>> I'd be in favor of making it a bloody law that every bloody language
>> use the same bloody spelling. I'm forever forgetting whether a
>> particular language uses ELSE IF, ELSEIF, ELSIF, or ELIF. Grumble,
>> grumble, grumble....

>
> As a relative newbie to PostgreSQL (but an old-timer to programming
> languages and other DMBSs) I would certainly vote for allowing elseif.
> This is my first encounter of "else" without the terminating "e", and
> that would not be a natural omission for me.
>
> --
> Guy Rouillier
>
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 9: the planner will ignore your desire to choose an index scan if
> your
> joining column's datatypes do not match
>
>

-----------------------------------------------------------
Frank D. Engel, Jr. <fde101@fjrhome.net>

$ ln -s /usr/share/kjvbible /usr/manual
$ true | cat /usr/manual | grep "John 3:16"
John 3:16 For God so loved the world, that he gave his only begotten
Son, that whosoever believeth in him should not perish, but have
everlasting life.
$



__________________________________________________ _________
$0 Web Hosting with up to 120MB web space, 1000 MB Transfer
10 Personalized POP and Web E-mail Accounts, and much more.
Signup at [url]www.doteasy.com[/url]


---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

[Clodoaldo Pinto]

--- Tom Lane <tgl@sss.pgh.pa.us> escreveu:

> Neil Conway <neilc@samurai.com> writes:

> > Tom also suggested just adding 'elseif' as an alternative for 'elsif'.
> > That sounds like it would be worth doing.

>
> I think we should go ahead and do that for 8.0. I'm getting tired of
> reading reports that stem from this mistake (I think this is the third
> one in the past month ...). I can't see any real downside to accepting
> both spellings, can you?
>
> regards, tom lane

I made the exact same mistake and it took me a good chunk of energy to figure
it out. I didn't report it to the list. When one look at the manual it is very
easy to read elseif in instead of elsif.

Regards,
Clodoaldo Pinto





__________________________________________________ _____
Yahoo! Mail - Agora com 250MB de espaço gratuito. Abra
uma conta agora! [url]http://br.info.mail.yahoo.com/[/url]

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [email]majordomo@postgresql.org[/email])

[Martijn van Oosterhout]

On Thu, Dec 16, 2004 at 05:14:42PM -0300, Clodoaldo Pinto wrote:

> --- Tom Lane <tgl@sss.pgh.pa.us> escreveu:

> > Neil Conway <neilc@samurai.com> writes:

> > > Tom also suggested just adding 'elseif' as an alternative for 'elsif'..
> > > That sounds like it would be worth doing.

> >
> > I think we should go ahead and do that for 8.0. I'm getting tired of
> > reading reports that stem from this mistake (I think this is the third
> > one in the past month ...). I can't see any real downside to accepting
> > both spellings, can you?
> >
> > regards, tom lane

>
> I made the exact same mistake and it took me a good chunk of energy to figure
> it out. I didn't report it to the list. When one look at the manual it isvery
> easy to read elseif in instead of elsif.

I'd vote for one of two options:

1. Accept both elseif and elsif as equivalent or
2. Add some parsing magic to give a meaningful error message so people
see it immediatly.

Maybe just match in "elseif" at beginning of a line and error out with
"maybe you meant elsif".
--
Martijn van Oosterhout <kleptog@svana.org> [url]http://svana.org/kleptog/[/url]

> Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
> tool for doing 5% of the work and then sitting around waiting for someone
> else to do the other 95% so you can sue them.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see [url]http://www.gnupg.org[/url]

iD8DBQFBwe8yY5Twig3Ge+YRAnvSAKCElzx0Jt/sE5YO7MqqFr7jSGR/FgCghrmu
8YO9rHg/eiEKvZHTZeYkYW4=
=FZ0F
-----END PGP SIGNATURE-----

[Richard_D_Levine@raytheon.com]

I prefer the FORTRAN66 construct

IF <condition> <label1> <label2> <label3>

where it jumps to label1 if <condition> is negative, label2 if zero, and
label3 if positive. No else ifs about it.

I hope you realize I'm as kidding as I am obviously too old.



Michael Fuhr
<mike@fuhr.org> To: Geoffrey <esoteric@3times25.net>
Sent by: cc: [email]pgsql-general@postgresql.org[/email]
pgsql-general-owner@pos Subject: Re: [GENERAL] pl/pgsql oddity
tgresql.org


12/16/2004 01:16 PM






On Thu, Dec 16, 2004 at 12:27:53PM -0500, Geoffrey wrote:

> I don't know of any other language that permits multiple spellings for
> the same construct. I'd be concerned with starting such a precedent.

I'd be in favor of making it a bloody law that every bloody language
use the same bloody spelling. I'm forever forgetting whether a
particular language uses ELSE IF, ELSEIF, ELSIF, or ELIF. Grumble,
grumble, grumble....

--
Michael Fuhr
[url]http://www.fuhr.org/~mfuhr/[/url]

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [email]majordomo@postgresql.org[/email] so that your
message can get through to the mailing list cleanly




---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match