From alexeys9 at mail.ru  Wed Dec 10 05:02:14 2008
From: alexeys9 at mail.ru (Makarov Alexey)
Date: Wed, 10 Dec 2008 10:02:14 +0500
Subject: [Cl-perec-devel] Problem with one-to-many association type
Message-ID: <1228885334.6637.27.camel@makarov>

Hello!

I have two classes:

(defpclass employee ()
  ((first-name :initarg :name :type string :accessor name-of)
   (second-name :initarg :second-name type string :accessor
second-name-of)
   (third-name :initarg :third-name :type string :accessor
third-name-of)))

(defpclass project ()
  ((code :initarg :code :type string :accessor code-of)
   (short-name :initarg :short-name :type string :accessor name-of)
   (full-name :initarg :full-name :type string :accessor full-name-of)))

I define association (one-to-many), and can't work with this classes any
more because of error:

(defassociation
  ((:class project :slot members :type (set employee) :accessor
members-of :initarg :members)
   (:class employee :slot project :type project :accessor
project-of :initarg :project)))

query:

(with-transaction
    (select (o)
        (from (o project))))

appear this error:

There is no applicable method for the generic function
  #<STANDARD-GENERIC-FUNCTION (SETF CL-RDBMS::UNIQUE-P) (1)>
when called with arguments
  (T NIL).
   [Condition of type SIMPLE-ERROR]

I'm use fresh cl-rdbms and fresh cl-perec.
How to solve this error and use one-to-many association?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-perec-devel/attachments/20081210/abe67bce/attachment.html>

From tomi.borbely at gmail.com  Wed Dec 10 08:23:14 2008
From: tomi.borbely at gmail.com (=?ISO-8859-1?Q?Borb=E9ly_Tam=E1s_?=)
Date: Wed, 10 Dec 2008 09:23:14 +0100
Subject: [Cl-perec-devel] Problem with one-to-many association type
In-Reply-To: <1228885334.6637.27.camel@makarov>
References: <1228885334.6637.27.camel@makarov>
Message-ID: <53c1332f0812100023l2f50d05ey839c5629a16a5ce0@mail.gmail.com>

It works for me. Stacktrace?
Removing the .fasl files before loading cl-perec might help.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-perec-devel/attachments/20081210/041a04f4/attachment.html>

From alexeys9 at mail.ru  Wed Dec 10 10:26:53 2008
From: alexeys9 at mail.ru (Makarov Alexey)
Date: Wed, 10 Dec 2008 15:26:53 +0500
Subject: [Cl-perec-devel] Problem with one-to-many association type
Message-ID: <1228904813.15226.1.camel@makarov>

>It works for me. Stacktrace?
>Removing the .fasl files before loading cl-perec might help.

I tried remove fasls, but error appear as before.
Stacktrace here.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-perec-devel/attachments/20081210/49b9c488/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: backtrace.tar.bz2
Type: application/x-bzip-compressed-tar
Size: 1267 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/cl-perec-devel/attachments/20081210/49b9c488/attachment.bin>

From levente.meszaros at gmail.com  Wed Dec 10 10:04:05 2008
From: levente.meszaros at gmail.com (=?ISO-8859-1?Q?Levente_M=E9sz=E1ros?=)
Date: Wed, 10 Dec 2008 11:04:05 +0100
Subject: [Cl-perec-devel] Problem with one-to-many association type
In-Reply-To: <1228885334.6637.27.camel@makarov>
References: <1228885334.6637.27.camel@makarov>
Message-ID: <a0d87c1c0812100204n2dc7c8f8kb3b01b626035f1d7@mail.gmail.com>

2008/12/10 Makarov Alexey <alexeys9 at mail.ru>:
>
> There is no applicable method for the generic function
>   #<STANDARD-GENERIC-FUNCTION (SETF CL-RDBMS::UNIQUE-P) (1)>
> when called with arguments
>   (T NIL).
>    [Condition of type SIMPLE-ERROR]
Some of your systems must be out of date, the exact code you have sent
runs without problems.
Double check that cl-perec dependencies are up to date (especially
cl-rdbms and computed-class).
Remove all fasl files.
If none of the above helps, give us more details of the error.

PRCT> (defpclass employee ()
        ((first-name :initarg :name :type string :accessor name-of)
         (second-name :initarg :second-name type string :accessor
second-name-of)
         (third-name :initarg :third-name :type string :accessor
third-name-of)))

#<PERSISTENT-CLASS EMPLOYEE>

PRCT> (defpclass project ()
  ((code :initarg :code :type string :accessor code-of)
   (short-name :initarg :short-name :type string :accessor name-of)
   (full-name :initarg :full-name :type string :accessor full-name-of)))

#<PERSISTENT-CLASS PROJECT>
PRCT> (defassociation
  ((:class project :slot members :type (set employee) :accessor
members-of :initarg :members)
   (:class employee :slot project :type project :accessor project-of
:initarg :project)))

PROJECT~MEMBERS~EMPLOYEE~PROJECT

PRCT> (start-sql-recording)
; No value

PRCT> (with-transaction
    (select (o)
        (from (o project))))

; BEGIN
; SELECT _o._oid, _o._code, _o._short_name, _o._full_name FROM _project_ap _o
; COMMIT
NIL

levy

-- 
There's no perfectoin



From levente.meszaros at gmail.com  Wed Dec 10 11:01:12 2008
From: levente.meszaros at gmail.com (=?ISO-8859-1?Q?Levente_M=E9sz=E1ros?=)
Date: Wed, 10 Dec 2008 12:01:12 +0100
Subject: [Cl-perec-devel] Problem with one-to-many association type
In-Reply-To: <1228904813.15226.1.camel@makarov>
References: <1228904813.15226.1.camel@makarov>
Message-ID: <a0d87c1c0812100301o298ed958o11af4de4d0db6b5f@mail.gmail.com>

2008/12/10 Makarov Alexey <alexeys9 at mail.ru>:
>>It works for me. Stacktrace?
>>Removing the .fasl files before loading cl-perec might help.
>
> I tried remove fasls, but error appear as before.
> Stacktrace here.
If something weird is happening then it might be a package problem.
>From the stacktrace it turned out that the association you have
defined is not a one to many association.
The reason for that is the symbol 'set' in the association definition
is not the one available from cl-perec and thus the type is unknown to
the type system. There must certainly be a good error message for this
case, I will fix that.

The solution is that you make sure that the symbol cl-perec:set is
used in your defassociation form, ie. don't shadow this in your
package or qualify explicitly.

levy

-- 
There's no perfectoin



From alexeys9 at mail.ru  Wed Dec 10 11:37:05 2008
From: alexeys9 at mail.ru (Makarov Alexey)
Date: Wed, 10 Dec 2008 16:37:05 +0500
Subject: [Cl-perec-devel] Problem with one-to-many association type
Message-ID: <1228909025.16920.3.camel@makarov>

Thank you! That's get worked! :)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-perec-devel/attachments/20081210/cd0e6ec0/attachment.html>

From alexeys9 at mail.ru  Fri Dec 12 19:10:02 2008
From: alexeys9 at mail.ru (Makarov Alexey)
Date: Sat, 13 Dec 2008 00:10:02 +0500
Subject: [Cl-perec-devel] Idea to query language
Message-ID: <1229109002.7457.10.camel@reginleif.ufanet.ru>

Hello.
I have an idea about query language.
Often I write conditions for querying objects that contains in their
slots (objects have the same class) one of values containing in the
list, and form "from" of query turn into long form "or" (or (eq slot 1)
(eq slot 2) ...), i create function that construct this condition:

(defun in (slot-name values)
`(or ,@(mapcar (lambda (v) `(eq (quote ,slot-name) ,v)) values)))

Then I can write:

(eval `(with-transaction
     (select (p)
       (from (o rank))
       (where ,(in 'name '("one" "two" "three"))))))

in PostgreSQL query language this presented as:

SELECT *
FROM class_table
WHERE slot IN VALUES (1::int,2,3)

http://www.postgresql.org/docs/8.3/interactive/sql-values.html

In databases that query languages support this syntax construction,
transform this as is,
in databases that query language not support this syntax, extract to
long form "OR".

As I know this feature included in SQL standard.

Maybe this would be worth to include as form in cl-perec query language?
We will can be use it without quote and outer "eval".
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-perec-devel/attachments/20081213/14a1370f/attachment.html>

From levente.meszaros at gmail.com  Fri Dec 12 19:56:10 2008
From: levente.meszaros at gmail.com (=?ISO-8859-1?Q?Levente_M=E9sz=E1ros?=)
Date: Fri, 12 Dec 2008 20:56:10 +0100
Subject: [Cl-perec-devel] Idea to query language
In-Reply-To: <1229109002.7457.10.camel@reginleif.ufanet.ru>
References: <1229109002.7457.10.camel@reginleif.ufanet.ru>
Message-ID: <a0d87c1c0812121156t1bcc4cpeb5947ed5cc20f0e@mail.gmail.com>

2008/12/12 Makarov Alexey <alexeys9 at mail.ru>:
> Hello.
> I have an idea about query language.
> Often I write conditions for querying objects that contains in their slots
> (objects have the same class) one of values containing in the list, and form--

I'm not sure, but is this what you really want?

PRCT> (defpclass* xxx ()
        ((slot :type integer-16)))
#<PERSISTENT-CLASS XXX>
PRCT> (with-transaction
        (make-instance 'xxx :slot 1))
#<XXX :persistent #t 66263>
PRCT> (start-sql-recording)
; No value
PRCT> (with-transaction
        (select (i)
          (from (i xxx))
          (where (member (slot-of i) '(1 2 3)))))
; BEGIN
; SELECT _i._oid, _i._slot FROM _xxx_ap _i WHERE ((_i._slot IN (1, 2, 3)))
; COMMIT
(#<XXX :persistent #t 66263>)
PRCT> (with-transaction
        (select (i)
          (from (i xxx))
          (where (member (slot-of i) '(10 20 30)))))
; BEGIN
; SELECT _i._oid, _i._slot FROM _xxx_ap _i WHERE ((_i._slot IN (10, 20, 30)))
; COMMIT
NIL

levy

There's no perfectoin



From alexeys9 at mail.ru  Sat Dec 13 07:11:10 2008
From: alexeys9 at mail.ru (=?koi8-r?Q?=E1=CC=C5=CB=D3=C5=CA_=ED=C1=CB=C1=D2=CF=D7?=)
Date: Sat, 13 Dec 2008 10:11:10 +0300
Subject: [Cl-perec-devel] =?koi8-r?b?SWRlYSB0byBxdWVyeSBsYW5ndWFnZQ==?=
In-Reply-To: <a0d87c1c0812121156t1bcc4cpeb5947ed5cc20f0e@mail.gmail.com>
References: <a0d87c1c0812121156t1bcc4cpeb5947ed5cc20f0e@mail.gmail.com>
Message-ID: <mailman.0.1428527118.30403.cl-perec-devel@common-lisp.net>

Yes, this is I want, thanks!