Dymanic cast

It is basically used for the safe type casting of base and derived classes.
Type casting from base class pointer to a derived class pointer is safe and does not require explicit casting, however reverse is not that much simple.
Type casting is required when we are pointing a base class object by a derived class pointer.
dynamic_cast is used for this type of herirachical casting ( upcasting ).
One pre requisite with this cast is - The base class should be polymorphic i.e. base class must have at least one virtual method.

In below example - Comment and comment out the virtual method of class A and check out the results.

#include "stdafx.h"
#include


class A
{
public :
A()
{
cout<<"A::A()"<}
~A()
{
cout<<"A::~A()"<}
//void virtual Test() {};
};


class B :public A
{
public :
B()
{
cout<<"B::B()"<}
~B()
{
cout<<"B::~B()"<}
};
int main(int argc, char* argv[])
{
try
{
A *objA = new A();
B *objB = new B();
objB = dynamic_cast(objA);
}
catch(...)
{
cout<<"error";
}
return 0;
}

Table Name Search for a Column

There could be situations when we need to know all the table names in our database which contain a required column name.The below stored procedure could be of use -
If exists ( select * from sysobjects where name

like 'GetTableName' and type = 'P'

)

BEGIN

DROP procedure GetTableName

END

GO

Create procedure GetTableName

@strString varchar(100)

As

Begin

Select name from sysobjects

where id in ( Select id from syscolumns where name like @strString )

End

GO

REALIZATION

In object oriented programming (OOPS), REALIZATION is a kind of relationship between two classes where one class provides the template for all the functions to be used , while the other class actually provides the functionalities to the defined functions.

The class providing the template is known as “Supplier” class while the class providing the implementation is called as “Consumer” class.

This type of relationship is used in Interface based architecture or template based architecture.

In the given example the class “Account” is an interface class and it gets implemented by the class
“SavingAccount”.

In UML the same thing is denoted as

#include "stdafx.h"


// Supplier class or Interface class
class Account
{
private:


protected:
float fPrincipal, fInterest, fTime, fRate;
public:
virtual void SetPrincipal(float fPrincipal)=0;
virtual void SetTime(float fTime)=0;
virtual void SetRate(float fRate)=0;
virtual void CalculateInterest()=0;
virtual void ShowInterest()=0;
};


// Consumer class or Implementor class
class SavingAccount : public Account
{
private:


protected:

public:
void SetPrincipal(float fPrincipal);
void SetTime(float fTime);
void SetRate(float fRate);
void CalculateInterest();
void ShowInterest();

};

void SavingAccount::SetPrincipal(float fPrincipal)
{
this->fPrincipal = fPrincipal;
}

void SavingAccount::SetTime(float fTime)
{
this->fTime = fTime;
}

void SavingAccount::SetRate(float fRate)
{
this->fRate = fRate;
}

void SavingAccount::CalculateInterest()
{
fInterest = (fPrincipal * fTime * fRate ) / 100;
}

void SavingAccount::ShowInterest()
{
printf("Intereset is - %f", fInterest);
}

int main(int argc, char* argv[])
{
SavingAccount objSavAcc;

objSavAcc.SetPrincipal(10000);
objSavAcc.SetRate(5.9);
objSavAcc.SetTime(2);
objSavAcc.ShowInterest();

return 0;
}

Clustered and non clustered index

Just to explain the main difference between Clustered and non clustered index with a simple example.

Clustered index makes data storage in a sorted order based on the key/s used for indexing, while non clustered index does not.

Just have a look to the below piece of T-SQL statements to remember this -


Clustered Index Example -


drop table Clustered_Table
Go

create table Clustered_Table
(
id int,

name varchar(100)
)
Go


--Create Clsutered Index
Create Clustered Index Clust_ind_id
on Clustered_Table(id)
GO


-- Insert few records
Insert into Clustered_Table Values (1, 'Avaneesh')
GO
Insert into Clustered_Table Values (5, 'Suresh')
GO
Insert into Clustered_Table Values (3, 'Dinesh')
GO
Insert into Clustered_Table Values (2, 'Mahesh')
GO
Insert into Clustered_Table Values (4, 'Vikash')
GO


O/P –


Id name
1 Avaneesh
2 Mahesh
3 Dinesh
4 Vikash
5 Suresh



Note the output is in sorted order.





Non Clustered Index Example -

drop table Non_Clustered_Table
Go
create table Non_Clustered_Table
(
id int,

name varchar(100)
)
Go


-- Create non clustered index
Create nonClustered Index test_Clust_ind1
on Non_Clustered_Table (id)
GO


-- Insert the records
Insert into Non_Clustered_Table Values (1, 'Avaneesh')
GO
Insert into Non_Clustered_Table Values (5, 'Suresh')
GO
Insert into Non_Clustered_Table Values (3, 'Dinesh')
GO
Insert into Non_Clustered_Table Values (2, 'Mahesh')
GO
Insert into Non_Clustered_Table Values (4, 'Vikash')
GO


-- get the records with non clustred index used
select * from Non_Clustered_Table



O/P –

Id Name
1 Avaneesh
5 Suresh
3 Dinesh
2 Mahesh
4 Vikash


Note the order of the records. Same as we had entered.

Hope it will help you somewhere.
Bye.

Façade pattern

Design patterns provide solutions to frequently occurring problems and also these patterns make the software efficient by suggesting effective ways.
FACADE is one the patterns which is commonly used.

What is Façade pattern ?
In this pattern the external access to a software system is made organized and efficient. A system containing no of sub systems (modules) can exposes many interfaces to the
external systems (clients ). Frequently for getting access to the required module, intelligence needs to be added at client side only. It requires internal knowledge of the
system being accessed and could make the client side little thick.
Facade pattern makes the system to expose only single interface i.e. point of contact and internally makes the decision about the redirection of the client request to the
correct module.

Advantages?
Thus the system access is more organized and easy for the client system.

The below example shows a Facade implementation. In this, there could be no of schemes in account system which have different formulae for interest rates. Depending upon the passed scheme code, correct scheme function is invoked to compute the interest rate.


#include "stdafx.h"
#include "iostream.h"

enum Scheme { A, B} ;

class SchemeA
{
public :

double GetInterest(float dbAmt, int intTenure)
{
return ( 0.02 * dbAmt / intTenure );
}
};


class SchemeB
{
public :

double GetInterest(float dbAmt, int intTenure)
{
return ( 0.03 * dbAmt / intTenure );
}
};

// Facade class with interface GetInterest(...)

class Facade
{
public:

double GetInterest(Scheme SchemeType, float dbAmt, int intTenure)
{

// decision making for reditrection of the client request
if (SchemeType == A)
{
SchemeA objA;

return (objA.GetInterest(dbAmt, intTenure));
}

if (SchemeType == B)
{
SchemeA objB;

return (objB.GetInterest( dbAmt, intTenure));
}

}
};


// client system for the above Facaded system
int main(int argc, char* argv[])
{
Facade objFacade;

// call to the Facade interface
cout<< objFacade.GetInterest(A, 5000, 14);
return 0;
}

Query Execution Plan

Fig 1

Fig 2

Often in database programming, we face the issue of query optimization. Though the Database Management System (DBMS) has one component known as "Query Optimizer" which takes care of optimization but programmer himself needs to improve the query to get the results efficiently.
While query optimization, there should be some benchmark / indicator to show the efficiency of the query which can help the programmer to decide which part to look on for optimization.

SQL Server 2000 has one such feature know as "Query Execution Plan".
One more terminology needs to be explained before going into that.

Cost of query - It figures how much CPU and I/O the query utilises. An optimized query should take less cost. Using this factor queries are compared for efficiency.

Query Execution Plan -
It means the way the SQL Server runs our query, how it optimizes and what are the costs for different actions of a query.

One very simple example -

There are two tables -Test1 and Test2 with 2 columns id and name each.
Selection the ids and name from these 2 table using following query -

Select T1.id, T2.Name
from test1 T1, test2 T2
where
T1.id = T2.id

Using the Query execution Plan we can do analysis of the costs occurred for the actions inside the query -

Check Fig 1.
It says -

1. Table scan 40 % cost - For scanning the tables T1 and T2, it costs 40 % of the entire query.

2. Hash Matching - For matching the fetched records as per the where clause it takes
19 %.

3. Select 0 % - Finally selecting the records.

(Where does the remaining 1 % go? )


Lets add something to the query - Say sorting -

Select T1.id, T2.Name
from test1 T1, test2 T2
where
T1.id = T2.id
order by T1.id

Check Fig 2.
Now the result has -

Scanning - 36 % each

Hash matching - 17 %

Sorting - This is new and because of the order by clause. Takes 11 %.

Finally select action.

(Its correct as 100 % total appears)

That’s all for this from me. Now your suggestion and advice.


Bye.
Avaneesh S. Tiwari
26 June '08

Dot net COM interoperability

"COM interoperability" a topic something important for a person coming from C++ / COM background to dot net world. Also a frequent question in dot net tech interviews -

I tried putting few test applications and knowing what it is.
In COM world the final output which we get is in the form of Typelibrary, a binary output.
While in dot net supported languages like C#, the output is know as Assembly, not binary but known as Intermediate Language (IL) output.

On a broad level, COM interoperability is transformation of these two different types of output files.
There could be 2 cases -

1. COM outputs to be used in dot net languages - We need to convert the TypeLibrary to assembly. It’s done by utility - TlbImp.exe - Type library import.

I never used this explicitly as there is simple option tab available while giving reference to COM dll in the dot net editor. It internally manages these all stuff of
Conversion.


2. Dot net dlls to be used in unmanaged environments - It requires conversion of Assembly to TypeLib. It could be done with - RegAsm.exe utility which stands for Register
assembly as Typelib. Honestly saying, I never required and used this.

This is what I could write, please share your experience.

Bye.
Avaneesh Tiwari

15-06-2008

Am beginner for non WINDOWS

After working for 4 years in WINDOWS world, I have got chance to work on non windows platform. With these tough days, as a beginner, I could share this much for now.

GNU -
GNU is an organization mainly in the area of Open Source software. Richard Stallman started it. It has been well known for the developments and distribution of tools, utilities, Applications etc. used in conjunction with Unix OS. GNU softwares are distributed with General Public License (GPL). Few applications / tools known to me -

1. GCC - Compiler used for languages like - C, C++, Objective C etc. Now the latest version is - 4.3.1. I am working on little older - 3.2.2 (February 05 release). It supports many languages as well as hardware platforms. With several setting, it can be customized and made to suit our business.

2. GDB - Debugger tool. With command line, it appeared tough for me to use as have migrated from Windows!

3. DDD - Something for my rescue. It uses GDB debugger and provides user interface to some extent. So its a user interface debugger for UNIX / Linux programmers.

Thats all for this article, please share your views.

-- By Avaneesh Tiwari

A good begining

Friends,

Its a lovely day as I am starting my technical blog. I believe this journey will be a successful one.

-- Avaneesh Tiwari

05-06-2008