From dpbowe at hotmail.com  Thu Feb  2 11:55:58 2017
From: dpbowe at hotmail.com (david bowe)
Date: Thu Feb  2 14:30:31 2017
Subject: [odb-users] query_one on postgres
Message-ID: <HE1PR08MB11613809C78858981F05A168C44C0@HE1PR08MB1161.eurprd08.prod.outlook.com>

Hi there,


I'm new to odb but I'm trying to upgrade my code with it to add an extra bit of flexibility.


I'm using Postgres (9.2.18) with odb (2.3.0), but the compiler is telling me that odb::database has no member query_one. See below


driver.cxx:71:13: error: 'std::auto_ptr<odb::database>::element_type' has no member named 'query_one'

         db->query_one<restaurants> (query::code


Could anybody provide any guidance on this issue as I have not been able to find anything on it so far


Thanks,


David
From boris at codesynthesis.com  Fri Feb  3 03:51:55 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Feb  3 03:52:04 2017
Subject: [odb-users] query_one on postgres
In-Reply-To: <HE1PR08MB11613809C78858981F05A168C44C0@HE1PR08MB1161.eurprd08.prod.outlook.com>
References: <HE1PR08MB11613809C78858981F05A168C44C0@HE1PR08MB1161.eurprd08.prod.outlook.com>
Message-ID: <boris.20170203105053@codesynthesis.com>

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> I'm using Postgres (9.2.18) with odb (2.3.0), but the compiler is
> telling me that odb::database has no member query_one.

That's because query_one() was added in ODB 2.4.0. See the NEWS file
in the ODB compiler package for details.

Boris

From dpbowe at hotmail.com  Fri Feb  3 03:53:56 2017
From: dpbowe at hotmail.com (david bowe)
Date: Fri Feb  3 08:56:28 2017
Subject: [odb-users] query_one on postgres
In-Reply-To: <boris.20170203105053@codesynthesis.com>
References: <HE1PR08MB11613809C78858981F05A168C44C0@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<boris.20170203105053@codesynthesis.com>
Message-ID: <HE1PR08MB1161DB218EF1D6916B352AA1C44F0@HE1PR08MB1161.eurprd08.prod.outlook.com>

Just realised that after I sent message
Thanks for the reply


> On 3 Feb 2017, at 08:52, Boris Kolpackov <boris@codesynthesis.com> wrote:
> 
> Hi David,
> 
> david bowe <dpbowe@hotmail.com> writes:
> 
>> I'm using Postgres (9.2.18) with odb (2.3.0), but the compiler is
>> telling me that odb::database has no member query_one.
> 
> That's because query_one() was added in ODB 2.4.0. See the NEWS file
> in the ODB compiler package for details.
> 
> Boris

From dpbowe at hotmail.com  Fri Feb  3 06:35:37 2017
From: dpbowe at hotmail.com (david bowe)
Date: Fri Feb  3 08:56:28 2017
Subject: [odb-users] query_one on postgres
In-Reply-To: <boris.20170203105053@codesynthesis.com>
References: <HE1PR08MB11613809C78858981F05A168C44C0@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<boris.20170203105053@codesynthesis.com>
Message-ID: <HE1PR08MB116108D5C44DDCC32002F007C44F0@HE1PR08MB1161.eurprd08.prod.outlook.com>

Hi Boris,


I found this question posted on the users group about a year ago:

http://www.codesynthesis.com/pipermail/odb-users/2016-February/003123.html


I am receiving a similar error when I run query_one:


driver: /usr/include/odb/result.txx:20: typename odb::object_traits<T>::pointer_type odb::result<T>::one() [with T = restaurants; typename odb::object_traits<T>::pointer_type = restaurants*]: Assertion `++i == end ()' failed.


Could you please advise me on the issue?

Also, since I'm already asking, with the #pragma db id auto unsigned long id_;
Should this automatically create public accessor functions or have I misinterpreted the docs?

Thanks,

David

________________________________
From: Boris Kolpackov <boris@codesynthesis.com>
Sent: 03 February 2017 08:51
To: david bowe
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] query_one on postgres

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> I'm using Postgres (9.2.18) with odb (2.3.0), but the compiler is
> telling me that odb::database has no member query_one.

That's because query_one() was added in ODB 2.4.0. See the NEWS file
in the ODB compiler package for details.

Boris
From xzhan156 at ford.com  Fri Feb  3 09:56:27 2017
From: xzhan156 at ford.com (Zhang, Xiaoguang (X.))
Date: Fri Feb  3 11:46:17 2017
Subject: [odb-users] A question about odb view
Message-ID: <aac78230eae8454f892b3684be8426e2@nafcmb40.exchange.ford.com>

Hi,
This is Jack Zhang. I am from Ford Motor NA. I discovered odb when I searched ORM software. I used it in my application which is part of our interval utilities. It worked very well and I like it very much.
Recently when I started using view and I found the view I created is not working for me. I hope you can point me out where is the miscoding or something else missing.
I am using SQLite of odb. I created a file like:

#ifndef __APPS_HXX__
#define __APPS_HXX__

#include <string>
#include <cstddef>

#include <odb/core.hxx>

#pragma db object
class apps
{
public:
    apps(const int appid, std::string &desc) : appid_(appid), desc_(desc) {}

    const int appid () { return appid_; }
    const std::string & desc () const { return desc_; }

    void appid (int aid ) {
        appid_ = aid;
    }
    void desc (const std::string & d) {
        desc_ = d;
    }

private:
    friend class odb::access;

    apps () {}

    #pragma db id auto
    unsigned long id_;

    #pragma db unique
    int appid_;

    std::string desc_;
};

#pragma db view object(apps)
struct apps_count
{
    #pragma db column ("count(" + apps::appid_ + ")")
    std::size_t count;
};

#endif

And this is the way I use it:

    typedef odb::query<apps_count> cquery;
typedef odb::result<apps_count> cresult;

    apps_count ac(odb_hdlr->query_value<apps_count> (cquery::appid == 1));
    cout << "apps record number: " << ac.count << endl;
    if (!ac.count) {
         cout << "There is no record for table ID: " << myedit_appids[i].id << ", creating one now ..." << endl;
           //string s(myedit_appids[i].desc);
            apps a(myedit_appids[i].id, (std::string &)myedit_appids[i].desc);
            //apps a(myedit_appids[i].id, s);
            odb_hdlr->persist(a);
}

NOTE:
Record with appid 1 is in database. Everything ac.count got zero back.

Odb_hdlr gets initialized like this, and it is successfully created without error and exceptions:
    int c = 3;
    char *v[4];
    v[0] = strdup("myedit");
    v[1] = strdup("--database");
    v[2] = strdup(adacs_dbname);
    v[3] = NULL;
odb_hdlr = new odb::sqlite::database (c, v, false, SQLITE_OPEN_READWRITE | SQLITE_OPEN_CREATE);

Thanks in advance. Looking forward for your answer!
Jack
From boris at codesynthesis.com  Mon Feb  6 07:17:26 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Feb  6 07:17:35 2017
Subject: [odb-users] query_one on postgres
In-Reply-To: <HE1PR08MB116108D5C44DDCC32002F007C44F0@HE1PR08MB1161.eurprd08.prod.outlook.com>
References: <HE1PR08MB11613809C78858981F05A168C44C0@HE1PR08MB1161.eurprd08.prod.outlook.com>
	<boris.20170203105053@codesynthesis.com>
	<HE1PR08MB116108D5C44DDCC32002F007C44F0@HE1PR08MB1161.eurprd08.prod.outlook.com>
Message-ID: <boris.20170206141458@codesynthesis.com>

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> driver: /usr/include/odb/result.txx:20:
> typename odb::object_traits<T>::pointer_type odb::result<T>::one()
> [with T = restaurants; typename odb::object_traits<T>::pointer_type =
> restaurants*]: Assertion `++i == end ()' failed.

Are you certain the query returns a single row?


> Also, since I'm already asking, with the #pragma db id auto unsigned
> long id_; Should this automatically create public accessor functions
> or have I misinterpreted the docs?

Yes, you have. ODB does not provide such functionality (since it has
nothing to do with ORM).

Boris

From boris at codesynthesis.com  Mon Feb  6 07:25:25 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Feb  6 07:25:34 2017
Subject: [odb-users] A question about odb view
In-Reply-To: <aac78230eae8454f892b3684be8426e2@nafcmb40.exchange.ford.com>
References: <aac78230eae8454f892b3684be8426e2@nafcmb40.exchange.ford.com>
Message-ID: <boris.20170206142245@codesynthesis.com>

Hi,

Zhang, Xiaoguang (X.) <xzhan156@ford.com> writes:

> Recently when I started using view and I found the view I created is
> not working for me. I hope you can point me out where is the miscoding
> or something else missing.

I don't see anything immediately obvious in your code. The hello example
that comes with ODB has a similar view and it is known to work so you may
want to try that. Alternatively, if you believe this is a bug in ODB, I
will need a minimal but complete test that reproduces the problem.

Boris

From marco.craveiro at gmail.com  Mon Feb  6 11:19:22 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Tue Feb  7 11:38:17 2017
Subject: [odb-users] Using odb with oracle: stored procedures and cursors
Message-ID: <CAKRNd4zhDTFNEkfG1oUeqfeDzAPewxjeTFNDD1Bb7Wg6pY8e5w@mail.gmail.com>

Hi odb-users,

Many thanks for a fantastic product. I am about to embark on a project
using ODB against an Oracle database, so I've started reading the
documentation. However, I could not find any references to stored
procedures or cursors in the Oracle section of the manual - although I
can see it for SQL Server and others. I also quickly googled the
mailinglist and grepped the source without success, so I thought I'd
ask. Having said that, I must confess it wasn't the most thorough of
searches - so I apologise in advance if I missed something obvious.

My questions are:

1. Is it possible at present to call stored procedures for Oracle?
2. Is it possible for stored procedures from Oracle to return multiple cursors?
3. If one of both of these features are not currently available, would
this require a lot of changes on the ODB side? I would consider
submitting a patch if the work is straightforward.

Also, I would consider gluing some hand-crafted code if required in
order to get this to work, provided it is straightforward - for
example, call the proc manually via the low-level ODB oracle API,
obtain the cursors and then supply each cursor to ODB generated code.

Many thanks for your time.
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key

From boris at codesynthesis.com  Tue Feb  7 12:14:43 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Feb  7 12:14:53 2017
Subject: [odb-users] Using odb with oracle: stored procedures and cursors
In-Reply-To: <CAKRNd4zhDTFNEkfG1oUeqfeDzAPewxjeTFNDD1Bb7Wg6pY8e5w@mail.gmail.com>
References: <CAKRNd4zhDTFNEkfG1oUeqfeDzAPewxjeTFNDD1Bb7Wg6pY8e5w@mail.gmail.com>
Message-ID: <boris.20170207190036@codesynthesis.com>

Hi Marco,

Marco Craveiro <marco.craveiro@gmail.com> writes:

> Many thanks for a fantastic product.

Thanks, I am glad you like it.


> However, I could not find any references to stored procedures or
> cursors in the Oracle section of the manual - although I can see
> it for SQL Server and others.

Right, we test/add support for stored procedures based on demand.


> 1. Is it possible at present to call stored procedures for Oracle?

We haven't got a test for it yet but it could already be working if
Oracle returns the result in a way similar to a SELECT query (and
all indications point to this being the case). If so, then calling
a procedure would require a view along these lines:

#pragma db view query("CALL my_stored_procedure(?)")
struct my_stored_procedure
{
  ...
};

And then:

using query = odb::query<my_stored_procedure>;

db.query<my_stored_procedure> (query::_val (123) + "," +
                               query::_val ("abc"));

I could whip up a test if you are interested.


> 2. Is it possible for stored procedures from Oracle to return multiple
>    cursors?

You mean multiple result sets (e.g., sets of rows from multiple,
potentially different, SELECT statements)? If so, that would most
likely require extra, potentially non-trivial, support.


> 3. If one of both of these features are not currently available, would
>    this require a lot of changes on the ODB side?

It's hard to say without further investigation, but achieving anything
with OCI (which is what libodb-oracle uses) is never straightforward.

Boris

From dpbowe at hotmail.com  Thu Feb  9 05:53:45 2017
From: dpbowe at hotmail.com (david bowe)
Date: Thu Feb  9 11:35:31 2017
Subject: [odb-users] Defining one-to-many bidirectional relationship
Message-ID: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>

Hi there,


I have been trying to define a one to many relationship with the attached code. However, when defining the relationship pointers I get the below error asking me to define a database type. I can do this and the code will compile but it creates a relationship through a supplier_order table which is not what I need. I need each order to reference one suppler


database.hxx:111:46: error: unable to map C++ type '::__gnu_cxx::new_allocator< ::boost::weak_ptr< ::order > >::this_type' used in data member 'orders_' to a PostgreSQL database type

database.hxx:111:46: info: use '#pragma db value_type' to specify the database type


Could anybody help me with what I am doing wrong?


Thanks,


David
-------------- next part --------------
// database.hxx
//

#ifndef DATABASE_HXX
#define DATABASE_HXX

#include <string>
#include <vector>
#include <odb/core.hxx>
#include <boost/shared_ptr.hpp>
#include <boost/weak_ptr.hpp>
#include <boost/date_time/posix_time/posix_time.hpp>

// Forward declaration
class order;

#pragma db object table("suppliers")
class supplier
{
public:
  typedef ::order orders_type;

  supplier (const std::string& title,
            const std::string& phone,
            const std::string& address_line_1,
            const std::string& address_line_2,
            const std::string& address_line_3,
            const std::string& address_line_4,
            const std::string& address_line_5)
    : title_ (title), phone_ (phone), 
      address_line_1_ (address_line_1), address_line_2_ (address_line_2),
      address_line_3_ (address_line_3), address_line_4_ (address_line_4),
      address_line_5_ (address_line_5)
  {}

  unsigned long id() const {
    return id_;
  }
  const std::string& title() const {
    return title_;
  }
  const std::string& phone() const {
    return phone_;
  }
  const std::string& address_line_1() const {
    return address_line_1_;
  }
  const std::string& address_line_2() const {
    return address_line_2_;
  }
  const std::string& address_line_3() const {
    return address_line_3_;
  }
  const std::string& address_line_4() const {
    return address_line_4_;
  }
  const std::string& address_line_5() const {
    return address_line_5_;
  }

  void title(const std::string& title) {
    title_ = title;
  }
  void phone(const std::string& phone) {
    phone_ = phone;
  }
  void address_line_1(const std::string& address_line_1) {
    address_line_1_ = address_line_1;
  }
  void address_line_2(const std::string& address_line_2) {
    address_line_2_ = address_line_2;
  }
  void address_line_3(const std::string& address_line_3) {
    address_line_3_ = address_line_3;
  }
  void address_line_4(const std::string& address_line_4) {
    address_line_4_ = address_line_4;
  }
  void address_line_5(const std::string& address_line_5) {
    address_line_5_ = address_line_5;
  }

private:
  friend class odb::access;

  supplier() {}

  #pragma db id auto
  unsigned long id_;

  #pragma db type("VARCHAR(255)") column("title")
  std::string title_;
  #pragma db type("VARCHAR(255)") column("phone") null
  std::string phone_;
  #pragma db type("VARCHAR(255)") column("address_line_1")
  std::string address_line_1_;
  #pragma db type("VARCHAR(255)") column("address_line_2") null
  std::string address_line_2_;
  #pragma db type("VARCHAR(255)") column("address_line_3") null
  std::string address_line_3_;
  #pragma db type("VARCHAR(255)") column("address_line_4") null
  std::string address_line_4_;
  #pragma db type("VARCHAR(255)") column("address_line_5") null
  std::string address_line_5_;
  #pragma db type("TIMESTAMP") column("created_at") options("DEFAULT CURRENT_TIMESTAMP")
  boost::posix_time::ptime created_;
  #pragma db type("TIMESTAMP") column("updated_at") options("DEFAULT CURRENT_TIMESTAMP")
  boost::posix_time::ptime updated_;

  #pragma db value_not_null inverse(supplier_) 
  std::vector<boost::weak_ptr<orders_type> > orders_;
};

#pragma db object table("orders")
class order
{
public:
  typedef ::supplier supplier_type;

  order (boost::shared_ptr<supplier_type> supplier,
         std::string bd_code)

    : supplier_ (supplier), bd_code_ (bd_code)
  {}

  unsigned long id() const {
    return id_;
  }
  const std::string& bd_code() const {
    return bd_code_;
  }

  void bd_code(const std::string& bd_code) {
    bd_code_ = bd_code;
  }

public:
  // Belongs to: supplier
  const boost::shared_ptr<supplier_type>& supplier() const {
    return supplier_;
  }

  void supplier(const boost::shared_ptr<supplier_type>& supplier) {
    supplier_ = supplier;
  }

private:
  friend class odb::access;

  order() {}

  #pragma db id auto
  unsigned long id_;

  #pragma db not_null type("VARCHAR(255)") column("code")
  std::string bd_code_;
  #pragma db type("TIMESTAMP") column("created_at") options("DEFAULT CURRENT_TIMESTAMP")
  boost::posix_time::ptime created_;
  #pragma db type("TIMESTAMP") column("updated_at") options("DEFAULT CURRENT_TIMESTAMP")
  boost::posix_time::ptime updated_;

  #pragma db not_null
  boost::shared_ptr<supplier_type> supplier_;

};

#endif // DATABASE_HXX

From marco.craveiro at gmail.com  Fri Feb 10 03:50:00 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Fri Feb 10 03:50:13 2017
Subject: [odb-users] Using odb with oracle: stored procedures and cursors
In-Reply-To: <boris.20170207190036@codesynthesis.com>
References: <CAKRNd4zhDTFNEkfG1oUeqfeDzAPewxjeTFNDD1Bb7Wg6pY8e5w@mail.gmail.com>
	<boris.20170207190036@codesynthesis.com>
Message-ID: <CAKRNd4wV8Vspk_w1NmUDz9k5am48pJTbAse-AyFOLOR34qraeA@mail.gmail.com>

Hi Boris,

thanks for the prompt reply.

<stored procs in oracle>
> We haven't got a test for it yet but it could already be working if
> Oracle returns the result in a way similar to a SELECT query (and
> all indications point to this being the case). If so, then calling
> a procedure would require a view along these lines:
>
> #pragma db view query("CALL my_stored_procedure(?)")
> struct my_stored_procedure
> {
>   ...
> };
>
> And then:
>
> using query = odb::query<my_stored_procedure>;
>
> db.query<my_stored_procedure> (query::_val (123) + "," +
>                                query::_val ("abc"));
>
> I could whip up a test if you are interested.

actually don't worry just yet - I changed the code to select the table
directly and that is working quite well. If we end up having to use
procs, I will use this to update the code and test it and give you
guys feedback.

> You mean multiple result sets (e.g., sets of rows from multiple,
> potentially different, SELECT statements)? If so, that would most
> likely require extra, potentially non-trivial, support.

Yes, unfortunately that is exactly what the procs are currently doing!
However, I will try to bypass these procs altogether as they don't
seem to fit the ODB/ORM model.

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key

From marco.craveiro at gmail.com  Fri Feb 10 03:59:37 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Fri Feb 10 03:59:51 2017
Subject: [odb-users] Using prefetch with Oracle
Message-ID: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>

Hi odb-users,

I have managed to do a successful test using ODB against an Oracle
database, and the ORM mapping part is working quite well. However, the
one snag I have is that fetching at the moment is a tad slow. It was
pointed out that I needed to enable prefetching in the Oracle drivers
so I've been investigating how to go about it. I can't see any direct
mentions in the manual; I saw references to bulk loading for
updates/deletes/inserts but not selects. However, I think I'm missing
something as I found this comment[1]:

> For fetches network traffic can be minimized using pre-fetching. ODB enables
> this feature if the underlying database supports it (e.g., Oracle).

This is exactly what I am looking for. But I grepped the code and
found no references to OCI_ATTR_PREFETCH_ROWS[2], so I tried manually
hacking "execute" as follows (statement.cxx in odb-oracle):

      const int prefetchSize(5000);
      sword r = OCIAttrSet (stmt_,
          OCI_HTYPE_STMT,
          (void*)&prefetchSize,
          sizeof(int),
          OCI_ATTR_PREFETCH_ROWS,
          err);

      if (r == OCI_ERROR || r == OCI_INVALID_HANDLE)
          translate_error (err, r);

Largely copied from SO[3]. This did not crash :-) but it didn't make
any performance difference either.

So in conclusion, any ideas on how to enable prefetching / bulk
loading for Oracle, using ODB? If this is not presently supported, how
should one extend the code to support it?

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key

[1] https://www.reddit.com/r/cpp/comments/p4hiq/odb_c_orm_180_released_adds_support_for_sql/
[2] https://docs.oracle.com/cd/B13789_01/appdev.101/b10779/oci04sql.htm
[3] http://stackoverflow.com/questions/26301386/call-ocistmtexecute-with-ref-cursor-and-oci-attr-prefetch-rows

From dpbowe at hotmail.com  Fri Feb 10 03:45:59 2017
From: dpbowe at hotmail.com (david bowe)
Date: Fri Feb 10 06:01:46 2017
Subject: [odb-users] Defining one-to-many bidirectional relationship
In-Reply-To: <adbe584d-060e-4572-9e8d-141c2084537f@email.android.com>
References: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<adbe584d-060e-4572-9e8d-141c2084537f@email.android.com>
Message-ID: <HE1PR08MB1161F9BC0575D96D61ED88A1C4440@HE1PR08MB1161.eurprd08.prod.outlook.com>

Thanks for the reply Micheal,


Although, I'm not quite sure I understand. Do you mean that I should only use a reference pointer on one side of the database map? I want to be able to query up and down through the relationship as necessary.


David

________________________________
From: Michael <mwpowellhtx@gmail.com>
Sent: 09 February 2017 17:29
To: david bowe
Subject: Re: [odb-users] Defining one-to-many bidirectional relationship



On February 9, 2017 5:53:45 AM EST, david bowe <dpbowe@hotmail.com> wrote:
>Hi there,
>
>
>I have been trying to define a one to many relationship with the
>attached code. However, when defining the relationship pointers I get
>the below error asking me to define a database type. I can do this and
>the code will compile but it creates a relationship through a
>supplier_order table which is not what I need. I need each order to
>reference one suppler

That's not typically how best to normalize this kind of object to database mapping. Doesn't matter if we're talking about odb/C++ or NHibernate/C#. What you described is more of a domain layer and/or business logic concern IMPO.

In this case the * on the order to supply side is 1, but the mapping can be the same.

HTH

>database.hxx:111:46: error: unable to map C++ type
>'::__gnu_cxx::new_allocator< ::boost::weak_ptr< ::order > >::this_type'
>used in data member 'orders_' to a PostgreSQL database type
>
>database.hxx:111:46: info: use '#pragma db value_type' to specify the
>database type
>
>
>Could anybody help me with what I am doing wrong?
>
>
>Thanks,
>
>
>David

--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
From dpbowe at hotmail.com  Fri Feb 10 10:07:18 2017
From: dpbowe at hotmail.com (david bowe)
Date: Fri Feb 10 10:38:06 2017
Subject: [odb-users] Defining one-to-many bidirectional relationship
In-Reply-To: <c51530e3-2fad-4cd0-a223-5f765932d1f5@email.android.com>
References: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<adbe584d-060e-4572-9e8d-141c2084537f@email.android.com>
	<HE1PR08MB1161F9BC0575D96D61ED88A1C4440@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<c51530e3-2fad-4cd0-a223-5f765932d1f5@email.android.com>
Message-ID: <HE1PR08MB1161CA39CE5CBBCC2FCD711FC4440@HE1PR08MB1161.eurprd08.prod.outlook.com>

Thanks Micheal,


I am creating a simple database schema so I will remove the inverse pointer. I think I misunderstood what is required for the relationship to be bidirectional (sorry still pretty new to this).


When I compile the code now it creates the correct schema (below), however, the supplier foreign key in the orders table does not show a reference to the suppliers table. Any ideas why this might be? The use of boost pointers maybe?? Sorry to keep asking


David


CREATE TABLE "suppliers" (

  "id" BIGSERIAL NOT NULL PRIMARY KEY,

  "title" VARCHAR(255) NOT NULL,

  "phone" VARCHAR(255) NULL,

  "address_line_1" VARCHAR(255) NOT NULL,

  "address_line_2" VARCHAR(255) NULL,

  "address_line_3" VARCHAR(255) NULL,

  "address_line_4" VARCHAR(255) NULL,

  "address_line_5" VARCHAR(255) NULL,

  "created_at" TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,

  "updated_at" TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP);


CREATE TABLE "orders" (

  "id" BIGSERIAL NOT NULL PRIMARY KEY,

  "code" VARCHAR(255) NOT NULL,

  "created_at" TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,

  "updated_at" TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,

  "supplier_id" BIGINT NOT NULL);




________________________________
From: Michael <mwpowellhtx@gmail.com>
Sent: 10 February 2017 14:48
To: david bowe
Subject: Re: [odb-users] Defining one-to-many bidirectional relationship



On February 10, 2017 3:45:59 AM EST, david bowe <dpbowe@hotmail.com> wrote:
>Thanks for the reply Micheal,
>
>
>Although, I'm not quite sure I understand. Do you mean that I should
>only use a reference pointer on one side of the database map? I want to
>be able to query up and down through the relationship as necessary.

I am not as familiar with the odb side of things as I am ORM, per se. AFAIK, *-* would support that, yes, with specific restrictions being domain concerns.

All I can encourage is just to think it through. Perhaps develop a couple of domain-only test cases to exercise the hypotheses.

HTH

>David
>
>________________________________
>From: Michael <mwpowellhtx@gmail.com>
>Sent: 09 February 2017 17:29
>To: david bowe
>Subject: Re: [odb-users] Defining one-to-many bidirectional
>relationship
>
>
>
>On February 9, 2017 5:53:45 AM EST, david bowe <dpbowe@hotmail.com>
>wrote:
>>Hi there,
>>
>>
>>I have been trying to define a one to many relationship with the
>>attached code. However, when defining the relationship pointers I get
>>the below error asking me to define a database type. I can do this and
>>the code will compile but it creates a relationship through a
>>supplier_order table which is not what I need. I need each order to
>>reference one suppler
>
>That's not typically how best to normalize this kind of object to
>database mapping. Doesn't matter if we're talking about odb/C++ or
>NHibernate/C#. What you described is more of a domain layer and/or
>business logic concern IMPO.
>
>In this case the * on the order to supply side is 1, but the mapping
>can be the same.
>
>HTH
>
>>database.hxx:111:46: error: unable to map C++ type
>>'::__gnu_cxx::new_allocator< ::boost::weak_ptr< ::order >
>>::this_type'
>>used in data member 'orders_' to a PostgreSQL database type
>>
>>database.hxx:111:46: info: use '#pragma db value_type' to specify the
>>database type
>>
>>
>>Could anybody help me with what I am doing wrong?
>>
>>
>>Thanks,
>>
>>
>>David
>
>--
>Sent from my Android device with K-9 Mail. Please excuse my brevity.

--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
From boris at codesynthesis.com  Fri Feb 10 10:45:20 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Feb 10 10:45:30 2017
Subject: [odb-users] Defining one-to-many bidirectional relationship
In-Reply-To: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>
References: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>
Message-ID: <boris.20170210174352@codesynthesis.com>

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> database.hxx:111:46: error: unable to map C++ type
> '::__gnu_cxx::new_allocator< ::boost::weak_ptr< ::order > >::this_type'
> used in data member 'orders_' to a PostgreSQL database type

You are using Boost smart pointers. Have you enabled the Boost profile
(libodb-boost)?

Boris

From boris at codesynthesis.com  Fri Feb 10 11:07:35 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Feb 10 11:07:44 2017
Subject: [odb-users] Using prefetch with Oracle
In-Reply-To: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>
References: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>
Message-ID: <boris.20170210180525@codesynthesis.com>

Hi Marco,

Marco Craveiro <marco.craveiro@gmail.com> writes:

> This is exactly what I am looking for. But I grepped the code and
> found no references to OCI_ATTR_PREFETCH_ROWS[2], so I tried manually
> hacking "execute" as follows (statement.cxx in odb-oracle):
> 
>       const int prefetchSize(5000);
>       sword r = OCIAttrSet (stmt_,
>           OCI_HTYPE_STMT,
>           (void*)&prefetchSize,
>           sizeof(int),
>           OCI_ATTR_PREFETCH_ROWS,
>           err);
> 
>       if (r == OCI_ERROR || r == OCI_INVALID_HANDLE)
>           translate_error (err, r);
> 
> Largely copied from SO[3]. This did not crash :-) but it didn't make
> any performance difference either.

Just to confirm, did you hack select_statement::execute()?

Boris

From dpbowe at hotmail.com  Fri Feb 10 11:33:00 2017
From: dpbowe at hotmail.com (david bowe)
Date: Sun Feb 12 07:25:14 2017
Subject: [odb-users] Defining one-to-many bidirectional relationship
In-Reply-To: <boris.20170210174352@codesynthesis.com>
References: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<boris.20170210174352@codesynthesis.com>
Message-ID: <HE1PR08MB11613FB5272BCDA12F98BA65C4440@HE1PR08MB1161.eurprd08.prod.outlook.com>

Hi Boris,


I have downloaded the libodb-boost profile and compiled it from source. Do i need to add any further options to the compiler or includes in the header file?


Thanks

________________________________
From: Boris Kolpackov <boris@codesynthesis.com>
Sent: 10 February 2017 15:45
To: david bowe
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] Defining one-to-many bidirectional relationship

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> database.hxx:111:46: error: unable to map C++ type
> '::__gnu_cxx::new_allocator< ::boost::weak_ptr< ::order > >::this_type'
> used in data member 'orders_' to a PostgreSQL database type

You are using Boost smart pointers. Have you enabled the Boost profile
(libodb-boost)?

Boris
From dpbowe at hotmail.com  Fri Feb 10 11:51:33 2017
From: dpbowe at hotmail.com (david bowe)
Date: Sun Feb 12 07:25:14 2017
Subject: [odb-users] Defining one-to-many bidirectional relationship
In-Reply-To: <boris.20170210174352@codesynthesis.com>
References: <HE1PR08MB11616B2ECF805B476E2D60C4C4450@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<boris.20170210174352@codesynthesis.com>
Message-ID: <HE1PR08MB1161D4D2EDB75A9CE1A6F296C4440@HE1PR08MB1161.eurprd08.prod.outlook.com>

Ok just saw in the docs that an option has to be added


$ odb -d pgsql --profile boost --generate-schema --generate-query database.hxx

________________________________
From: Boris Kolpackov <boris@codesynthesis.com>
Sent: 10 February 2017 15:45
To: david bowe
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] Defining one-to-many bidirectional relationship

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> database.hxx:111:46: error: unable to map C++ type
> '::__gnu_cxx::new_allocator< ::boost::weak_ptr< ::order > >::this_type'
> used in data member 'orders_' to a PostgreSQL database type

You are using Boost smart pointers. Have you enabled the Boost profile
(libodb-boost)?

Boris
From marco.craveiro at gmail.com  Mon Feb 13 05:25:38 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Mon Feb 13 05:25:52 2017
Subject: [odb-users] Using prefetch with Oracle
In-Reply-To: <boris.20170210180525@codesynthesis.com>
References: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>
	<boris.20170210180525@codesynthesis.com>
Message-ID: <CAKRNd4z7h7A9Dj5UHhbofcJncCeskowGcFgsN-+hqwnF8AA62Q@mail.gmail.com>

Boris,

<adding OCI_ATTR_PREFETCH_ROWS>
> Just to confirm, did you hack select_statement::execute()?

Aha! I must have changed it in the wrong place! :-) I re-did my change
as I had reverted it but this time I made sure it was on
select_statement::execute(); I can now see a massive difference in
performance (~5x faster cold). Quick hack attached.

I'm going to have a look and see if I can follow the same pattern for
the bulk parameter so that it uses the pragma value when setting the
prefetch (and does nothing if bulk is not set). If I do this, would
you accept it as a patch?

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key
-------------- next part --------------
A non-text attachment was scrubbed...
Name: prefetch.patch
Type: text/x-patch
Size: 1194 bytes
Desc: not available
Url : http://codesynthesis.com/pipermail/odb-users/attachments/20170213/63f637f3/prefetch.bin
From anaswara.nn at gmail.com  Mon Feb 13 06:05:48 2017
From: anaswara.nn at gmail.com (Anaswara Nair)
Date: Mon Feb 13 06:06:00 2017
Subject: [odb-users] Store JSON value to DB
Message-ID: <CAF6eBr0ptnfp7E6JBX1TcAWh-eo_5dYmBXS20-na=zt79kGW8Q@mail.gmail.com>

Is there any method other than by converting it to wstring to store a JSON
value as such to SQLite database using ODB?
From boris at codesynthesis.com  Mon Feb 13 08:36:53 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Feb 13 08:37:02 2017
Subject: [odb-users] Store JSON value to DB
In-Reply-To: <CAF6eBr0ptnfp7E6JBX1TcAWh-eo_5dYmBXS20-na=zt79kGW8Q@mail.gmail.com>
References: <CAF6eBr0ptnfp7E6JBX1TcAWh-eo_5dYmBXS20-na=zt79kGW8Q@mail.gmail.com>
Message-ID: <boris.20170213152203@codesynthesis.com>

Hi,

Anaswara Nair <anaswara.nn@gmail.com> writes:

> Is there any method other than by converting it to wstring to store a JSON
> value as such to SQLite database using ODB?

SQLite just added the json1 loadable extension[1]. While it still stores
the data as TEXT (as discussed in the documentation, due to backwards
compatibility that's the only way), it provides some functions that can
be useful.

If I were to use this extension with ODB I would probably map an
extended database type[2] so that I get automatic validation and
minification on INSERT/UPDATE:

#pragma db map type("JSON_TEXT") as("TEXT") to("json((?))")

#pragma db object
struct object
{
  #pragma db type("JSON_TEXT")
  std::string json;
};

And then use the other functions json1 provides in ODB views if I
wanted to extract only some parts of the JSON document.

[1] https://sqlite.org/json1.html
[2] http://www.codesynthesis.com/~boris/blog/2012/07/18/custom-database-to-cxx-type-mapping-in-odb/

Boris

From boris at codesynthesis.com  Mon Feb 13 08:56:52 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Feb 13 08:57:02 2017
Subject: [odb-users] Using prefetch with Oracle
In-Reply-To: <CAKRNd4z7h7A9Dj5UHhbofcJncCeskowGcFgsN-+hqwnF8AA62Q@mail.gmail.com>
References: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>
	<boris.20170210180525@codesynthesis.com>
	<CAKRNd4z7h7A9Dj5UHhbofcJncCeskowGcFgsN-+hqwnF8AA62Q@mail.gmail.com>
Message-ID: <20170213135652.GA22900@codesynthesis.com>

Hi Marco,

Marco Craveiro <marco.craveiro@gmail.com> writes:

> I'm going to have a look and see if I can follow the same pattern for
> the bulk parameter so that it uses the pragma value when setting the
> prefetch (and does nothing if bulk is not set). If I do this, would
> you accept it as a patch?

Hm, I am not sure the bulk approach (with a compiler-time pragma) is
right in this case. There we don't really have a choice since we need
to know the "batch buffer" size.

But here it is all runtime. Plus, you may want to have different
prefetch for different queries of the same object. In fact, you
can already customize it for queries (but not for object loads)
by using prepared queries (Section 4.5 in the manual):

1. Create prepared query.

2. Get its statement (statement()).

3. Cast it to odb::oracle::select_statement.

4. Call handle() on the result to get OCIStmt*.

5. Set custom OCI_ATTR_PREFETCH_ROWS.

6. Execute the query.

The problems with this approach are: (1) it is tedious and (2) it
doesn't work for non-query SELECT's (e.g., database::load()). So
perhaps the way to do it is:

1. Provide prefetch() functions on oracle::database() and
   oracle::connection() that can be used to modify database-wide
   and connection-wide prefetch values. Also set it to some
   reasonable default (say 512?)

2. Provide oracle::select_statement::prefetch() to make the
   prepared query approach less tedious.

What do you think?

Boris

From marco.craveiro at gmail.com  Mon Feb 13 09:00:03 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Mon Feb 13 09:00:17 2017
Subject: [odb-users] Using prefetch with Oracle
In-Reply-To: <20170213135652.GA22900@codesynthesis.com>
References: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>
	<boris.20170210180525@codesynthesis.com>
	<CAKRNd4z7h7A9Dj5UHhbofcJncCeskowGcFgsN-+hqwnF8AA62Q@mail.gmail.com>
	<20170213135652.GA22900@codesynthesis.com>
Message-ID: <CAKRNd4yoBmOsTe6D_ZRm1yAYz-9g-YCk75hunfP0OnW1JzMjMA@mail.gmail.com>

Hi Boris,

<prefetch hack>
> Hm, I am not sure the bulk approach (with a compiler-time pragma) is
> right in this case. There we don't really have a choice since we need
> to know the "batch buffer" size.
>
> But here it is all runtime. Plus, you may want to have different
> prefetch for different queries of the same object. In fact, you
> can already customize it for queries (but not for object loads)
> by using prepared queries (Section 4.5 in the manual):
>
> 1. Create prepared query.
>
> 2. Get its statement (statement()).
>
> 3. Cast it to odb::oracle::select_statement.
>
> 4. Call handle() on the result to get OCIStmt*.
>
> 5. Set custom OCI_ATTR_PREFETCH_ROWS.
>
> 6. Execute the query.
>
> The problems with this approach are: (1) it is tedious and (2) it
> doesn't work for non-query SELECT's (e.g., database::load()). So
> perhaps the way to do it is:
>
> 1. Provide prefetch() functions on oracle::database() and
>    oracle::connection() that can be used to modify database-wide
>    and connection-wide prefetch values. Also set it to some
>    reasonable default (say 512?)
>
> 2. Provide oracle::select_statement::prefetch() to make the
>    prepared query approach less tedious.
>
> What do you think?

Yep, makes perfect sense. Let me have a go at coding this then.

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key

From dpbowe at hotmail.com  Tue Feb 14 12:18:17 2017
From: dpbowe at hotmail.com (david bowe)
Date: Wed Feb 15 03:58:25 2017
Subject: [odb-users] Defining database with multiple table mapping files
Message-ID: <HE1PR08MB1161F771570FCBA7944704C0C4580@HE1PR08MB1161.eurprd08.prod.outlook.com>

Hi there,


In the given examples all object classes are defined within one file.


However, with a number of tables making up my database and want to define each in a separate header file for maintainability. Is there a best practice for doing this in terms of compilation?


At the moment I can either paste all headers into one file and pass it to the odb compiler, or try to compile each individually and include each into my database driver file. Could anybody advise what is best?


Thanks,


David
From boris at codesynthesis.com  Wed Feb 15 11:10:28 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Feb 15 11:10:38 2017
Subject: [odb-users] Defining database with multiple table mapping files
In-Reply-To: <HE1PR08MB1161F771570FCBA7944704C0C4580@HE1PR08MB1161.eurprd08.prod.outlook.com>
References: <HE1PR08MB1161F771570FCBA7944704C0C4580@HE1PR08MB1161.eurprd08.prod.outlook.com>
Message-ID: <boris.20170215180825@codesynthesis.com>

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> At the moment I can either paste all headers into one file and pass
> it to the odb compiler, or try to compile each individually and
> include each into my database driver file.

Individual compilation/inclusion is definitely the recommended approach.
And if you need to generate a single .sql file, then you can use the
--at-once ODB compiler option.

Boris

From xiang.zhao at gamegou.com  Thu Feb 16 02:51:15 2017
From: xiang.zhao at gamegou.com (=?utf-8?B?WmhhbyBYaWFuZw==?=)
Date: Thu Feb 16 11:32:31 2017
Subject: [odb-users] Atomic and batch update in ODB
Message-ID: <tencent_01C5BAD958F79F9660765B9D@qq.com>

Dear ODB Developers:


I've been using ODB and MySQL for quite some time, It's been very useful for my projects, and I think it would be much more helpful if the following features could be included in odb:


a) Atomic Update


Considering the following code from http://www.codesynthesis.com/products/odb/doc/manual.xhtml#3.5:


void
update_age (database& db, person& p)
{
  try
  {
    transaction t (db.begin ());


    p.age (p.age () + 1);
    db.update (p);


    t.commit ();
  }
  catch (...)
  {
    transaction t (db.begin ());
    db.load (p.id (), p);
    t.commit ();


    throw;
  }
}
  

If this function is called in two threads at the same time, it could result in only one age in added.


Is there any plan for feature that equivalent to the following SQL: update `person` set `age` = `age` + 1 where `id` = xxx; ?


b) Batch Update


For now I can do a query and update the results in a for-loop, but it would be much slower than a single update statement like: update `person` set `age` = `age` + 1 where `first_name` = 'Joe';


Or: update `person` set `age`= case `id` when 1 then 10 when 2 then 20 else `id` where id in (1, 2);


c) Batch Insert


Similar the previous one, it would be helpful to have something like this: insert into `person` (`id`, `age`) values (1, 10), (2, 20);


d) Insert and duplicate


This feature can be implemented with find() and insert() now, however it would be atomic and faster with something like this: insert into `person` (`id`, `age`) values (1, 10) on duplicate key ignore;


Or: insert into `person` (`id`, `age`) values (1, 10) on duplicate key update set `age` = values(`age`)


Or in some case: insert into `person` (`id`, `age`) values (1, 10) on duplicate key update set `age` += values(`age`)




I hope my suggestions could be taken into consideration in your further development plan.


Thanks
Zhao Xiang
From dpbowe at hotmail.com  Thu Feb 16 04:10:48 2017
From: dpbowe at hotmail.com (david bowe)
Date: Thu Feb 16 11:32:31 2017
Subject: [odb-users] Defining database with multiple table mapping files
In-Reply-To: <boris.20170215180825@codesynthesis.com>
References: <HE1PR08MB1161F771570FCBA7944704C0C4580@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<boris.20170215180825@codesynthesis.com>
Message-ID: <HE1PR08MB11613B77AAFF68C2E5CA23B0C45A0@HE1PR08MB1161.eurprd08.prod.outlook.com>

Thanks Boris,


This is great. So I can just include the the everthing-odb.cxx file generated by --at-once into my driver?

Also I appreciate that this isn't really an odb question, but how should I set the make file target for the odb compiler since it generates multiple output files?

Thanks,

David

________________________________
From: Boris Kolpackov <boris@codesynthesis.com>
Sent: 15 February 2017 16:10
To: david bowe
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] Defining database with multiple table mapping files

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> At the moment I can either paste all headers into one file and pass
> it to the odb compiler, or try to compile each individually and
> include each into my database driver file.

Individual compilation/inclusion is definitely the recommended approach.
And if you need to generate a single .sql file, then you can use the
--at-once ODB compiler option.

Boris
From dpbowe at hotmail.com  Thu Feb 16 05:09:58 2017
From: dpbowe at hotmail.com (david bowe)
Date: Thu Feb 16 11:32:31 2017
Subject: [odb-users] Defining database with multiple table mapping files
In-Reply-To: <HE1PR08MB11613B77AAFF68C2E5CA23B0C45A0@HE1PR08MB1161.eurprd08.prod.outlook.com>
References: <HE1PR08MB1161F771570FCBA7944704C0C4580@HE1PR08MB1161.eurprd08.prod.outlook.com>,
	<boris.20170215180825@codesynthesis.com>,
	<HE1PR08MB11613B77AAFF68C2E5CA23B0C45A0@HE1PR08MB1161.eurprd08.prod.outlook.com>
Message-ID: <HE1PR08MB1161A142EBDB8B1887FE51FCC45A0@HE1PR08MB1161.eurprd08.prod.outlook.com>

Ok sorry stupid question, I have this sorted now


________________________________
From: david bowe <dpbowe@hotmail.com>
Sent: 16 February 2017 09:10
To: odb-users@codesynthesis.com
Subject: Re: [odb-users] Defining database with multiple table mapping files


Thanks Boris,


This is great. So I can just include the the everthing-odb.cxx file generated by --at-once into my driver?

Also I appreciate that this isn't really an odb question, but how should I set the make file target for the odb compiler since it generates multiple output files?

Thanks,

David

________________________________
From: Boris Kolpackov <boris@codesynthesis.com>
Sent: 15 February 2017 16:10
To: david bowe
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] Defining database with multiple table mapping files

Hi David,

david bowe <dpbowe@hotmail.com> writes:

> At the moment I can either paste all headers into one file and pass
> it to the odb compiler, or try to compile each individually and
> include each into my database driver file.

Individual compilation/inclusion is definitely the recommended approach.
And if you need to generate a single .sql file, then you can use the
--at-once ODB compiler option.

Boris
From boris at codesynthesis.com  Thu Feb 16 11:50:25 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Feb 16 11:50:35 2017
Subject: [odb-users] Atomic and batch update in ODB
In-Reply-To: <tencent_01C5BAD958F79F9660765B9D@qq.com>
References: <tencent_01C5BAD958F79F9660765B9D@qq.com>
Message-ID: <boris.20170216184757@codesynthesis.com>

Hi Zhao,

Yes, we have this feature on our TODO list (we call it "mass update"
not to be confused with "batch update" as found in Oracle/SQL Server).
I've added a note about your email to this item.

In the meantime, you can probably implement some/most of this using
optimistic concurrency (to achieve atomicity) and/or native statements.

Thanks,
Boris

From harriev9 at gmail.com  Sun Feb 19 07:06:45 2017
From: harriev9 at gmail.com (harriev9)
Date: Sun Feb 19 07:07:19 2017
Subject: [odb-users] Compile libodb-qt
Message-ID: <1487506005.6682.5.camel@gmail.com>

Hi,

I have working on compiling libodb-qt on a fedora 25 installation.

I had trouble with running ./configure. It complaint about not finding
QtCore.

First I added the -fPIC option, but this did not help:

./configure CXXFLAGS=-fPIC

There was another problem. Qt uses #include <QtCore/QString> in its
headers and so also /usr/include/qt5 was needed in the include path.

At the end i found out that i had to use:

./configure CXXFLAGS=-fPIC??CPPFLAGS=-I/usr/include/qt5 LDFLAGS=-
L/usr/lib64/

Is this the correct way to solve the problem?

From boris at codesynthesis.com  Mon Feb 20 12:23:37 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Feb 20 12:23:46 2017
Subject: [odb-users] Compile libodb-qt
In-Reply-To: <1487506005.6682.5.camel@gmail.com>
References: <1487506005.6682.5.camel@gmail.com>
Message-ID: <boris.20170220192029@codesynthesis.com>

Hi,

harriev9 <harriev9@gmail.com> writes:

> First I added the -fPIC option, but this did not help:
> 
> ./configure CXXFLAGS=-fPIC

Yes, Qt now required that. We will look into adding this automatically
for the next release.


> There was another problem. Qt uses #include <QtCore/QString> in its
> headers and so also /usr/include/qt5 was needed in the include path.

This path should be tried automatically.


> At the end i found out that i had to use:
> 
> ./configure CXXFLAGS=-fPIC CPPFLAGS=-I/usr/include/qt5 LDFLAGS=-
> L/usr/lib64/
> 
> Is this the correct way to solve the problem?

That depends on what the problem was. Can you look into config.log
to see which tests were tried and why they are failing?

Boris

From gilles.benepougoue at gmail.com  Mon Feb 20 18:53:51 2017
From: gilles.benepougoue at gmail.com (Gilles Bene)
Date: Mon Feb 20 18:54:04 2017
Subject: [odb-users] Compilation error when calling
	QLazyWeakPointer<T>::load()
Message-ID: <CACmEO1K18=G1TsyttG0ZuqrJHhGotWBUH63KWf2drKini=TSQg@mail.gmail.com>

Hi .
Problem, when calling the load() methode of a QLazyWeakPointer please need
help.

I have a class with a member as a QLazyWeakPointer
example:

class SomeClass {

 //...
QLazyWeakPointer<Bill> bill_;

//....
}

qreal Bill::getAmount() const{
    return bill_.load()->Amount();
  }

When I compile I get this error message.

/usr/include/odb/qt/smart-ptr/lazy-ptr.ixx: In instantiation of
'QSharedPointer<T> QLazyWeakPointer<T>::load() const [with T = Sortie]':
Facture.cpp:28:18:   required from here
/usr/include/odb/qt/smart-ptr/lazy-ptr.ixx:473:7: error: no match for
'operator=' (operand types are 'QSharedPointer<Sortie>' and
'odb::object_traits<Sortie>::pointer_type {aka Sortie*}')
     r = i_.template load<T> (false); // Keep id.
       ^
I get the same error anytime I call the load() methode of a
QLazyWeakPointer (
)
Please help.
From boris at codesynthesis.com  Tue Feb 21 11:49:22 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Feb 21 11:49:30 2017
Subject: [odb-users] Compilation error when calling
	QLazyWeakPointer<T>::load()
In-Reply-To: <CACmEO1K18=G1TsyttG0ZuqrJHhGotWBUH63KWf2drKini=TSQg@mail.gmail.com>
References: <CACmEO1K18=G1TsyttG0ZuqrJHhGotWBUH63KWf2drKini=TSQg@mail.gmail.com>
Message-ID: <boris.20170221184350@codesynthesis.com>

Hi Gilles,

Gilles Bene <gilles.benepougoue@gmail.com> writes:

> /usr/include/odb/qt/smart-ptr/lazy-ptr.ixx:473:7: error: no match for
> 'operator=' (operand types are 'QSharedPointer<Sortie>' and
> 'odb::object_traits<Sortie>::pointer_type {aka Sortie*}')
>      r = i_.template load<T> (false); // Keep id.
>        ^

Well, the compiler tells you exactly what the problem is: load() returns
Sortie* and left-hand-side is QSharedPointer<Sortie>. And that assignment
is not valid. Which one would be valid? For example, QSharedPointer<Sortie>
to QSharedPointer<Sortie>. Why is load() returning Sortie* instead of
QSharedPointer<Sortie>? Because you haven't told ODB that your object
pointers should be something other than raw pointers (see Section 3.3,
"Object and View Pointers" in the manual).

Boris


From gilles.benepougoue at gmail.com  Wed Feb 22 13:45:55 2017
From: gilles.benepougoue at gmail.com (Gilles Bene)
Date: Wed Feb 22 13:46:08 2017
Subject: [odb-users] Compilation error when calling
	QLazyWeakPointer<T>::load()
In-Reply-To: <boris.20170221184350@codesynthesis.com>
References: <CACmEO1K18=G1TsyttG0ZuqrJHhGotWBUH63KWf2drKini=TSQg@mail.gmail.com>
	<boris.20170221184350@codesynthesis.com>
Message-ID: <CACmEO1JTJ+7mGFPiPp0wJM91b1qpGU=J8Ar4eZd9SnFG7P1D_Q@mail.gmail.com>

Hi,
Thanks for your prompt reply.
Actually I want load() to return QSharedPointer<Sortie> instead of Sortie*.

I've' used the --default-pointer option to specify the default pointer for
the
whole file by suppliying it as ODB Compiler Command Line option in my make
file.
And I can see that the option is actually supplied to odb at compile time
as I can see the output of make calling odb.

odb -x -fPIC -x --std=c++11 --database pgsql --sql-name-case lower
--profile qt --default-pointer QSharedPointer --hxx-prologue-file
../pointer-par-defaut.hxx --generate-query --generate-session
-I/usr/include/x86_64-linux-gnu/qt5 Sortie.hxx

the file  pointer-par-defaut.hxx contains the line:
#include <QtCore/QSharedPointer>

But I still have the same error message.
I've check the file Sortie-odb.hxx wether it actually include the
definition
of the default pointer. I notice that it actually has the include #include
<QtCore/QSharedPointer>
Here is a part of the file Sortie-odb.hxx:

// This file was generated by ODB, object-relational mapping (ORM)
// compiler for C++.
//

#ifndef SORTIE_ODB_HXX
#define SORTIE_ODB_HXX

// Begin prologue.
//
#include <odb/qt/version.hxx>
#if ODB_QT_VERSION != 2040000 // 2.4.0
#  error ODB and C++ compilers see different libodb-qt interface versions
#endif
#include <odb/qt/basic/pgsql/qstring-traits.hxx>
#include <odb/qt/basic/pgsql/qbyte-array-traits.hxx>
#include <odb/qt/basic/pgsql/quuid-traits.hxx>
#include <odb/qt/containers/qhash-traits.hxx>
#include <odb/qt/containers/qlist-traits.hxx>
#include <odb/qt/containers/qlinked-list-traits.hxx>
#include <odb/qt/containers/qmap-traits.hxx>
#include <odb/qt/containers/qset-traits.hxx>
#include <odb/qt/containers/qvector-traits.hxx>
#include <odb/qt/date-time/pgsql/qdate-traits.hxx>
#include <odb/qt/date-time/pgsql/qtime-traits.hxx>
#include <odb/qt/date-time/pgsql/qdate-time-traits.hxx>
#include <QtCore/QSharedPointer>
#include <odb/qt/smart-ptr/pointer-traits.hxx>
#include <odb/qt/smart-ptr/wrapper-traits.hxx>
#include <QtCore/QSharedPointer>


//
// End prologue.

#include <odb/version.hxx>

#if (ODB_VERSION != 20400UL)
#error ODB runtime version mismatch
#endif

#include <odb/pre.hxx>

#include "Sortie.hxx"

#include "../DocsRelances/Relancable-odb.hxx"
#include "Relance-odb.hxx"
#include "../Personnels/Personnel-odb.hxx"
#include "Emballage-odb.hxx"
#include "Fournisseur-odb.hxx"
#include "HistPrix-odb.hxx"
#include "PrixFourn-odb.hxx"
#include "../Produits/Produit-odb.hxx"
#include "Tonnage-odb.hxx"
#include "DetailsSortie-odb.hxx"
#include "Facture-odb.hxx"
#include "Reglement-odb.hxx"

#include <memory>
#include <cstddef>
#include <string>

#include <odb/core.hxx>
#include <odb/traits.hxx>
#include <odb/callback.hxx>
#include <odb/wrapper-traits.hxx>
#include <odb/pointer-traits.hxx>
#include <odb/container-traits.hxx>
#include <odb/session.hxx>
#include <odb/cache-traits.hxx>
#include <odb/polymorphic-info.hxx>
#include <odb/result.hxx>
#include <odb/polymorphic-object-result.hxx>

#include <odb/details/unused.hxx>
#include <odb/details/shared-ptr.hxx>

namespace odb
{
  // Sortie
  //
  template <>
  struct class_traits< ::Sortie >
  {
    static const class_kind kind = class_object;
  };

  template <>
  class access::object_traits< ::Sortie >
  {
    public:
    typedef ::Sortie object_type;
    typedef ::QSharedPointer< ::Sortie > pointer_type;
    typedef odb::pointer_traits<pointer_type> pointer_traits;

    static const bool polymorphic = true;

    typedef ::Sortie root_type;
    typedef ::std::string discriminator_type;
    typedef polymorphic_map<object_type> map_type;
    typedef polymorphic_concrete_info<object_type> info_type;

// ......

I tought the line
typedef ::QSharedPointer< ::Sortie > pointer_type;
Was the specifying the default pointer for all instances of Sortie.
I looked for anything like odb::object_traits<Sortie>::pointer_type
as appearing in the error message but couldn't fine.
May be am getting something wrong.
Please need your advice. Sorry for my lengthy message.

On 21 February 2017 at 17:49, Boris Kolpackov <boris@codesynthesis.com>
wrote:

> Hi Gilles,
>
> Gilles Bene <gilles.benepougoue@gmail.com> writes:
>
> > /usr/include/odb/qt/smart-ptr/lazy-ptr.ixx:473:7: error: no match for
> > 'operator=' (operand types are 'QSharedPointer<Sortie>' and
> > 'odb::object_traits<Sortie>::pointer_type {aka Sortie*}')
> >      r = i_.template load<T> (false); // Keep id.
> >        ^
>
> Well, the compiler tells you exactly what the problem is: load() returns
> Sortie* and left-hand-side is QSharedPointer<Sortie>. And that assignment
> is not valid. Which one would be valid? For example, QSharedPointer<Sortie>
> to QSharedPointer<Sortie>. Why is load() returning Sortie* instead of
> QSharedPointer<Sortie>? Because you haven't told ODB that your object
> pointers should be something other than raw pointers (see Section 3.3,
> "Object and View Pointers" in the manual).
>
> Boris
>
>
>


-- 
Gilles BENE POUGOUE
From dpbowe at hotmail.com  Thu Feb 23 04:34:40 2017
From: dpbowe at hotmail.com (david bowe)
Date: Thu Feb 23 11:27:42 2017
Subject: [odb-users] Custom tracer
Message-ID: <HE1PR08MB1161ED3EF9DC36D50AB16213C4530@HE1PR08MB1161.eurprd08.prod.outlook.com>

Hi there,


I am implementing a custom tracer to log odb querys in a file using the given example in the manual.


However, when adding the custom tracer through the db pointer I am told I am not able to access the tracer odb::tracer class. The compiler gives the following error.


src/db-driver.cxx:68:21: error: ?odb::tracer? is an inaccessible base of ?pgsql_tracer?

   db->tracer (tracer);

                     ^


Any help would be gratefully appreciated [?]

The following is my database connection function


void Database::connectDatabase() {


        const std::string username = "user";

        const std::string password = "pass";

        const std::string database = "database";

        const std::string host = "localhost";

        const unsigned int port = 5432;


        try {


                pgsql_tracer tracer;

                db.reset(new odb::pgsql::database (username, password, database, host, port));

                db->tracer (tracer);


        } catch (odb::exception& e) {


                FILE_LOG(logERROR) << e.what();

                exit(1);


        }

}


Thanks,


David
-------------- next part --------------
A non-text attachment was scrubbed...
Name: =?utf-8?B?T3V0bG9va0Vtb2ppLfCfmKEucG5n?=
Type: image/png
Size: 461 bytes
Desc: =?utf-8?B?T3V0bG9va0Vtb2ppLfCfmKEucG5n?=
Url : http://codesynthesis.com/pipermail/odb-users/attachments/20170223/1769f95f/utf-8BT3V0bG9va0Vtb2ppLfCfmKEucG5n.png
From boris at codesynthesis.com  Thu Feb 23 11:54:23 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Feb 23 11:54:33 2017
Subject: [odb-users] Compilation error when calling
	QLazyWeakPointer<T>::load()
In-Reply-To: <CACmEO1JTJ+7mGFPiPp0wJM91b1qpGU=J8Ar4eZd9SnFG7P1D_Q@mail.gmail.com>
References: <CACmEO1K18=G1TsyttG0ZuqrJHhGotWBUH63KWf2drKini=TSQg@mail.gmail.com>
	<boris.20170221184350@codesynthesis.com>
	<CACmEO1JTJ+7mGFPiPp0wJM91b1qpGU=J8Ar4eZd9SnFG7P1D_Q@mail.gmail.com>
Message-ID: <boris.20170223185031@codesynthesis.com>

Hi Gilles,

Gilles Bene <gilles.benepougoue@gmail.com> writes:

>   template <>
>   class access::object_traits< ::Sortie >
>   {
>     public:
>     typedef ::Sortie object_type;
>     typedef ::QSharedPointer< ::Sortie > pointer_type;

Yes, this looks correct.


> I looked for anything like odb::object_traits<Sortie>::pointer_type
> as appearing in the error message but couldn't fine.

Hm, do you have Sortie-odb.hxx included in the source file where you
call load()? odb::object_traits<Sortie> (defined in <odb/traits.hxx>)
derives from odb::access::object_traits<Sortie>.

Boris

From boris at codesynthesis.com  Thu Feb 23 12:02:22 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Feb 23 12:02:31 2017
Subject: [odb-users] Custom tracer
In-Reply-To: <HE1PR08MB1161ED3EF9DC36D50AB16213C4530@HE1PR08MB1161.eurprd08.prod.outlook.com>
References: <HE1PR08MB1161ED3EF9DC36D50AB16213C4530@HE1PR08MB1161.eurprd08.prod.outlook.com>
Message-ID: <boris.20170223185951@codesynthesis.com>

Hi David

david bowe <dpbowe@hotmail.com> writes:

> I am implementing a custom tracer to log odb querys in a file using
> the given example in the manual.
> 
> However, when adding the custom tracer through the db pointer I am told
> I am not able to access the tracer odb::tracer class. The compiler gives
> the following error.
> 
> src/db-driver.cxx:68:21: error: ?odb::tracer? is an inaccessible base
> of ?pgsql_tracer?
> 
>    db->tracer (tracer);

If only you read that section on tracers until the end, you would have
had your answer. Quoting Section 3.13, "Tracing SQL Statement Execution":

"Note also that you can only set a database-specific tracer object using a
database-specific database instance, for example:"

pgsql_tracer tracer;

odb::database& db = ...;
db.tracer (tracer); // Compile error.

odb::pgsql::database& db = ...;
db.tracer (tracer); // Ok.

Boris

From gilles.benepougoue at gmail.com  Thu Feb 23 12:44:53 2017
From: gilles.benepougoue at gmail.com (Gilles Bene)
Date: Thu Feb 23 12:45:07 2017
Subject: [odb-users] Compilation error when calling
	QLazyWeakPointer<T>::load()
In-Reply-To: <boris.20170223185031@codesynthesis.com>
References: <CACmEO1K18=G1TsyttG0ZuqrJHhGotWBUH63KWf2drKini=TSQg@mail.gmail.com>
	<boris.20170221184350@codesynthesis.com>
	<CACmEO1JTJ+7mGFPiPp0wJM91b1qpGU=J8Ar4eZd9SnFG7P1D_Q@mail.gmail.com>
	<boris.20170223185031@codesynthesis.com>
Message-ID: <CACmEO1KV3yb2=cdHSKzJQY4M1dFcJ+EshSc8SJRTVR+NcaBorw@mail.gmail.com>

Wouah , thanks a lot for your quality technical supply. I've included the
file
Sortie-odb.hxx Sortie-odb.hxx in the source file where I call load() and
every thing is ok.
Thanks once more. You have a fabulous peace of software. I've been using it
for just a month
and am loving it.

On 23 February 2017 at 17:54, Boris Kolpackov <boris@codesynthesis.com>
wrote:

> Hi Gilles,
>
> Gilles Bene <gilles.benepougoue@gmail.com> writes:
>
> >   template <>
> >   class access::object_traits< ::Sortie >
> >   {
> >     public:
> >     typedef ::Sortie object_type;
> >     typedef ::QSharedPointer< ::Sortie > pointer_type;
>
> Yes, this looks correct.
>
>
> > I looked for anything like odb::object_traits<Sortie>::pointer_type
> > as appearing in the error message but couldn't fine.
>
> Hm, do you have Sortie-odb.hxx included in the source file where you
> call load()? odb::object_traits<Sortie> (defined in <odb/traits.hxx>)
> derives from odb::access::object_traits<Sortie>.
>
> Boris
>



-- 
Gilles BENE POUGOUE
From marco.craveiro at gmail.com  Fri Feb 24 07:23:53 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Fri Feb 24 07:24:07 2017
Subject: [odb-users] Using prefetch with Oracle
In-Reply-To: <CAKRNd4yoBmOsTe6D_ZRm1yAYz-9g-YCk75hunfP0OnW1JzMjMA@mail.gmail.com>
References: <CAKRNd4x7K-WPp7C2xTtCfQ0rD4ou-4FYaxpEFeOU1vrHeWAbTA@mail.gmail.com>
	<boris.20170210180525@codesynthesis.com>
	<CAKRNd4z7h7A9Dj5UHhbofcJncCeskowGcFgsN-+hqwnF8AA62Q@mail.gmail.com>
	<20170213135652.GA22900@codesynthesis.com>
	<CAKRNd4yoBmOsTe6D_ZRm1yAYz-9g-YCk75hunfP0OnW1JzMjMA@mail.gmail.com>
Message-ID: <CAKRNd4xW6SPCpD9tZ-Pbw236B8dK4D3yMfd3pzKZdNHBdRQNzw@mail.gmail.com>

Hi Boris,

<prefetch chanes>
>> So perhaps the way to do it is:
>>
>> 1. Provide prefetch() functions on oracle::database() and
>>    oracle::connection() that can be used to modify database-wide
>>    and connection-wide prefetch values. Also set it to some
>>    reasonable default (say 512?)
>>
>> 2. Provide oracle::select_statement::prefetch() to make the
>>    prepared query approach less tedious.
>>
>> What do you think?
>
> Yep, makes perfect sense. Let me have a go at coding this then.

Just a quick heads up, I am working on this but I ended up having to
first replicate the setup at home so it's taking me quite a while. On
the plus side I am cataloguing the adventure under a series of blog
posts - [1], [2] so far - which may (or may not) help other Oracle /
ODB users. Should be getting to the code soon, hopefully.

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key

[1] Nerd Food: Northwind, or Using Dogen with ODB - Part I:
http://mcraveiro.blogspot.co.uk/2017/02/nerd-food-northwind-or-using-dogen-with.html
[2] Nerd Food: Northwind, or Using Dogen with ODB - Part II:
http://mcraveiro.blogspot.co.uk/2017/02/nerd-food-northwind-or-using-dogen-with_24.html

From harriev9 at gmail.com  Sun Feb 26 10:49:42 2017
From: harriev9 at gmail.com (harriev9)
Date: Sun Feb 26 10:49:57 2017
Subject: [odb-users] Compile libodb-qt
In-Reply-To: <boris.20170220192029@codesynthesis.com>
References: <1487506005.6682.5.camel@gmail.com>
	<boris.20170220192029@codesynthesis.com>
Message-ID: <1488124182.14426.1.camel@gmail.com>

Hi,

I am sorry. I must have done something wrong since i cant reproduce the
problem anymore.

I have another problem however trying to compile my code.
I constantly get an error:


/usr/local/include/odb/qt/basic/pgsql/default-mapping.hxx:8:26: fatal
error: QtCore/QString: No such file or directory


I think the include files are correct, during configuration i logged
the include directories:

-- !=!=! dir='/usr/include/qt5'
-- !=!=! dir='/usr/include/qt5/QtWidgets'
-- !=!=! dir='/usr/include/qt5/QtGui'
-- !=!=! dir='/usr/include/qt5/QtCore'
-- !=!=! dir='/usr/lib64/qt5/./mkspecs/linux-g++'
-- !=!=! dir='/usr/include'
-- !=!=! dir='/usr/include/qt5'
-- !=!=! dir='/usr/include/qt5/QtCore'
-- !=!=! dir='/usr/lib64/qt5/./mkspecs/linux-g++'
-- !=!=! dir='/usr/include/qt5'
-- !=!=! dir='/usr/include/qt5/QtGui'
-- !=!=! dir='/usr/include/qt5/QtCore'
-- !=!=! dir='/usr/lib64/qt5/./mkspecs/linux-g++'
-- !=!=! dir='/usr/include'
-- !=!=! dir='/home/harrie/projects/OpenControl/Components/src'
-- !=!=! dir='/home/harrie/projects/OpenControl/Components'
-- !=!=! dir='/home/harrie/projects/OpenControl/Components/src'
-- !=!=! 

What can be the problem here?

Harrie


On Mon, 2017-02-20 at 19:23 +0200, Boris Kolpackov wrote:
> Hi,
> 
> harriev9 <harriev9@gmail.com> writes:
> 
> > First I added the -fPIC option, but this did not help:
> > 
> > ./configure CXXFLAGS=-fPIC
> 
> Yes, Qt now required that. We will look into adding this
> automatically
> for the next release.
> 
> 
> > There was another problem. Qt uses #include <QtCore/QString> in its
> > headers and so also /usr/include/qt5 was needed in the include
> > path.
> 
> This path should be tried automatically.
> 
> 
> > At the end i found out that i had to use:
> > 
> > ./configure CXXFLAGS=-fPIC CPPFLAGS=-I/usr/include/qt5 LDFLAGS=-
> > L/usr/lib64/
> > 
> > Is this the correct way to solve the problem?
> 
> That depends on what the problem was. Can you look into config.log
> to see which tests were tried and why they are failing?
> 
> Boris

From harriev9 at gmail.com  Sun Feb 26 11:52:14 2017
From: harriev9 at gmail.com (harriev9)
Date: Sun Feb 26 11:52:28 2017
Subject: [odb-users] 
	[Fwd: Compile libodb-qt], Correction. Found the problem.
References: <1487506005.6682.5.camel@gmail.com>
Message-ID: <1488127934.14426.3.camel@gmail.com>

Hi,

I had a problem, the compiler complained QtCore/QString not found.

I found, after a long search, that the odb compiler needs the include
directories passed too.

I use cmake. So i added:

?INCLUDE ${Qt5Core_INCLUDE_DIRS}?

to:

odb_compile(...)

I also had to add?
  STANDARD c++11

since qt5 requires this.

Cheers,

Harrie

-------------- next part --------------
An embedded message was scrubbed...
From: harriev9 <harriev9@gmail.com>
Subject: Compile libodb-qt
Date: Sun, 19 Feb 2017 13:06:45 +0100
Size: 901
Url: http://codesynthesis.com/pipermail/odb-users/attachments/20170226/d2e41eb8/attachment.mht
From marco.craveiro at gmail.com  Sun Feb 26 18:03:25 2017
From: marco.craveiro at gmail.com (Marco Craveiro)
Date: Sun Feb 26 18:03:38 2017
Subject: [odb-users] BLOBs and multi-database
Message-ID: <CAKRNd4yVUQSzn4PgHVdkiB+0mx0SOF5HxrHTeDJR1FP-=Vs5eg@mail.gmail.com>

Hi odb-users,

I am using ODB with multiple databases (PostgreSQL and Oracle). A
couple of my fields is are BLOB/BYTEA. I have managed to map these
successfully to each of the databases individually via the pragma
type, but I don't seem to be able to map them to each database
"simultaneously" when on multi-database scenario.

I think my google-fu/documentation-fu is failing me badly - I am
convinced I had seen a way to set the type on a per-database basis but
for the life of me I just can't find that reference again.

Cheers
-- 
Marco Craveiro

MD, Domain Driven Consulting
about: http://about.me/marcocraveiro
blog: http://mcraveiro.blogspot.co.uk
twitter: https://twitter.com/MarcoCraveiro

That the Ideas are themselves manifestations (of the Idea-Idea) and
that the Idea-Idea is a-kind-of Manifestation-Idea?which is a-kind-of
itself, so that the system is completely self-describing? would have
been appreciated by Plato as an extremely practical joke [Plato]. --
Alan Key

From boris at codesynthesis.com  Mon Feb 27 07:04:21 2017
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Feb 27 07:04:29 2017
Subject: [odb-users] BLOBs and multi-database
In-Reply-To: <CAKRNd4yVUQSzn4PgHVdkiB+0mx0SOF5HxrHTeDJR1FP-=Vs5eg@mail.gmail.com>
References: <CAKRNd4yVUQSzn4PgHVdkiB+0mx0SOF5HxrHTeDJR1FP-=Vs5eg@mail.gmail.com>
Message-ID: <boris.20170227140051@codesynthesis.com>

Hi Marco,

Marco Craveiro <marco.craveiro@gmail.com> writes:

> [...] but I don't seem to be able to map them to each database
> "simultaneously" when on multi-database scenario.

You are probably looking for the database pragma prefix (introduction
to Chapter 16, "Multi-Database Support"):

#pragma db pgsql:type("BYTEA") oracle:type("BLOB")
std::vector<char> data;

Boris

From sean.clarke at sec-consulting.co.uk  Tue Feb 28 15:49:11 2017
From: sean.clarke at sec-consulting.co.uk (Sean Clarke)
Date: Tue Feb 28 15:49:24 2017
Subject: [odb-users] Inverse and composite Id
Message-ID: <CAHCKNJiT6aa8JH0udvO=8nRR58y1u5bnBfyHCJkfeX7YDDCE6A@mail.gmail.com>

Hi,
    I am having some problems trying to model a many-to-one relationship
where one entity has a composite primary key (I think that is the issue).

Trying to define what I am doing (this is a cut down definition - ignore
any compilation errors, this is just to summarise):


class A {

#pragma db value
    struct Id {
        int a;
        int b;
}

#pragma db id
    Id m_id;

#pragma db value_not_null
    std::shared_ptr<B> m_b;

};


class B {
    using id_t = std::uint32_t;

#pragma db id
    id_t m_id;

#pragma db inverse(m_b)
    std::vector<std::shared_ptr<A>> m_as;
};


The error I get when I compile is along the lines of:


/<path>/odb_gen/A-odb.hxx:385:70: error: ?image_type? in ?class
odb::access::composite_value_traits<B::Id, (odb::database_id)2u>? does not
name a type
         composite_value_traits< ::B::Id, id_pgsql >::image_type
value_value;

^~~~~~~~~~

Any hints on what I am doing wrong?


Regards
Sean Clarke