Srihari's blog

UDF for Splitting a varchar in SQL Server

noreply@blogger.com (Srihari Kothapalli) — Mon, 25 Aug 2008 16:24:00 +0000

set ANSI_NULLS ON
set QUOTED_IDENTIFIER ON
go

CREATE FUNCTION [dbo].[Split]
(
@RowData nvarchar(2000),
@SplitOn nvarchar(5)
)
RETURNS @RtnValue table
(
Id int identity(1,1),
Data nvarchar(100)
)
AS
BEGIN
Declare @Cnt int
Set @Cnt = 1
While (Charindex(@SplitOn,@RowData)>0)
Begin
Insert Into @RtnValue (data)
Select
Data = ltrim(rtrim(Substring(@RowData,1,Charindex(@SplitOn,@RowData)-1)))
Set @RowData = Substring(@RowData,Charindex(@SplitOn,@RowData)+1,len(@RowData))
Set @Cnt = @Cnt + 1
End
Insert Into @RtnValue (data)
Select Data = ltrim(rtrim(@RowData))
Return
END

Sending email through SSIS (Script task)

noreply@blogger.com (Srihari Kothapalli) — Tue, 19 Aug 2008 14:16:00 +0000

Sending email through VB.NET (Could be used in SSIS script task)

Dim mailMsg As System.Web.Mail.MailMessage
mailMsg = New System.Web.Mail.MailMessage()
mailMsg.From = "srihari@kothapalli.com"
mailMsg.To = "srihari.kothapalli@gmail.com"
mailMsg.Cc = ""
mailMsg.Bcc = ""
mailMsg.Subject = "This is subject at " + DateTime.Now.ToString()
mailMsg.BodyFormat = System.Web.Mail.MailFormat.Html
mailMsg.Body = "This is the body"
mailMsg.Priority = System.Web.Mail.MailPriority.High
System.Web.Mail.SmtpMail.SmtpServer = "smtp.kothapalli.com"
mailMsg.Fields.Add("http://schemas.microsoft.com/cdo/configuration/smtpauthenticate", "1")
mailMsg.Fields.Add("http://schemas.microsoft.com/cdo/configuration/sendusername", "kot\srihari")
mailMsg.Fields.Add("http://schemas.microsoft.com/cdo/configuration/sendpassword", "srihar1pa33w0rd")
mailMsg.Fields.Add("http://schemas.microsoft.com/cdo/configuration/smtpserverport", "25")
mailMsg.Fields.Add("http://schemas.microsoft.com/cdo/configuration/smtpusessl", "false")
System.Web.Mail.SmtpMail.Send(mailMsg)

Hope it helped.

Mingle tool by ThoughtWorks

noreply@blogger.com (Srihari Kothapalli) — Mon, 18 Aug 2008 13:15:00 +0000

Agile Project Management Tool

I was long wondering, how does the project management get automated in Agile practices with all the planning games being driven by a program instead of paper cards!!! Until I have seen one of my friends blogs – I did not realize that ThoghtWorks actually has a open source project management tool named “Mingle” to help this scenario.

Find his blog entry here and download the Mingle here. Download has taken me around 30 seconds. It is really a quick download and installation has taken me around 1 minute. So, its quite good with the features its promising to deliver.

It says it can not work with anything other than MySQL or PostGreSQL. It is a problem for me given that it is only SQL Server or Oracle that are the market leaders.

However, just to explore the tool.. lets go forward a bit more.

Downloaded the MySQL from here.

Followed the steps below.

1. Installed mySQL and used “create database mingle;” command in the command prompt of mySQL to create the database for setting up the Mingle on my localbox.
2. Ran through all the 6 steps in the http://localhost:8080/install/migrate as needed by Mingle.
3. Skipped the SMTP server details to fill in later

Not to forget, we have lot of help in understanding Mingle at http://studios.thoughtworks.com/mingle

When I started using this, as a website it looked quite good. Great thing about this is to be able to select the pre-defined template as below.

• Agile hybrid template(2.0.1)
This template represents a hybrid Agile approach to project tracking that would support release and iteration-based planning. The template includes some sample cards to illustrate the working of the charts and transitions. These stories can be deleted if necessary by a Project Administrator.
• Scrum template(2.0.1)
• Xp template(2.0.1)
This template can be used to create an Extreme Programming (XP) project for a small team that needs basic card tracking capability and reporting. The template contains some sample stories to illustrate how one might use charts and transitions. These stories can be deleted by a Project Administrator.
However, my research had to end here inspite of seeing all the really good tabs that says “sprint backlog” etc… Because most of the tabs keep giving me the error that says “error etc etc”.
I will give another try on my home box and let you know if I get any success.
But the summary is that today or tomorrow, it is going to get stabilized. It is a great news for Industry which is already behind agile approaches of software development.

My kid being loved by my mother

noreply@blogger.com (Srihari Kothapalli) — Sun, 17 Aug 2008 17:58:00 +0000

My brother started blogging !!!

noreply@blogger.com (Srihari Kothapalli) — Sat, 16 Aug 2008 12:38:00 +0000

Great news... My brother started blogging.. Reach him at

http://sureshkothapalli.blogspot.com

Job interview screening questions and answers

noreply@blogger.com (Srihari Kothapalli) — Fri, 15 Aug 2008 13:24:00 +0000

I was looking at my earlier mails and I found this bunch of questions and answers which I had to give for one of the interviews in London. Position was for a Lead Developer who would work in SCRUM/TDD practices of agile environment.

http://www.w3.org/TR/REC-html40">
Nokia Standard Document Template

Can you please describe the Red Green Refactor Processin context of test driven development?

Every TDDfeature development cycle goes through multiple iterations of Red, Green andRefactor.

The processsays that first we automate the test that we would like to add to our evolvingproduct/project. And then run the test till it succeeds without worrying aboutthe code/design. After we get the test to pass, we then start to enhance thecode quality by doing refactoring.

Red: We must run the testbefore we actually code the feature and the test is supposed to fail. This isto test the automated test.

Green: We would see Red untilwe see start seeing green which means that test passes.

Refactor: Once we get the test topass, then we start freely making changes to code to improve the design(refactoring). We are able to do this because we have the confidence ofautomated unit tests which would ensure that none of the existing featureschange because of the changes.

2. Can you pleasedescribe the singleton pattern?

Singleton isone of the creational patterns. Essentially, singleton is a class which allowsonly single instance to be created of itself. This is one patterns that couldbe tempting to use for many of the programming needs. And overuse simply killsthe performance. FULL STOP.

The approach Iusually take against this is not to use until it becomes mandatory.

Well, to startdrill into the details. We can discuss about two sides of the singleton. Oneface is about ways to create the singleton and the other face is to make surethat the singleton that we create would really stay as a singleton.

Ways tocreate

1. Creating astatic instance variable and creating the instance at the class load time

2. Usinglazy initialization with a static getInstance method

Both the approaches do not allow us to be able to override thegetinstance methods which would leave the polymorphic behaviour.

3. Factoryclass with a method which could return a singleton instance.

4.singletonfactory class that stores the mapping between the class names and itssingleton instances

Theseare more powerful approaches than the first two!!!

On theother hand, to make sure singleton is really a singleton... We need to be awareof quite a lot of things, I would point some of them here. (I am running shortof the space here.. will summarize this. I can explain it in much detailed waygiven an opportunity for the interview)

1. Serialize and deserialize the singleton object

2. Instantiatethe serialize the object in two or more app domains.

3. Garbagecollector could come into picture and might collect it since there is no bodyoutside of it is holding a reference.

Etc.

Usage

Some of thecommon uses include, having a logging service singleton class, to hold adatabase connection where we have limited number of connections that can bemade eg. With the Experian NBS copy database.

3. Can you tell ustell us the different types of joins available in SQL and describe them?

Basicallyjoins are used to get the subset of a Cartesian product from two tables.Broadly, there are two types of joins.
1. Inner join: A join of two tables which omits the rows if either sideis a NULL.

2. OuterJoin:Outer join is subcategorised in to Left Outer Join and Right Outer Join. Whichincludes blank rows in specified side if conditionsatisfies. Simple outer join iscombination of left and right outer joins.

Apart fromthese there are :

3. Naturaljoin:Just the Cartesian product
4. Equi join: which includes = operator in condition
5. Non Equi join: All conditional joins which doesn't uses = in theconditions.

4. Can you describe contract firstdevelopment and why we use it?

When creatingweb services, there are two different styles of developing the services.

· Contractfirst:C++ days, WCF is nearly there

· Codefirst:VB days, ASMX implementation

Whenwe are doing contract first, we start with the WSDL contract first and thencode for that contract. This helps in solving lot of problems includingthe below.

1. Fragility: No guarantee thatcontract stays the same over the long time

2. Performance is high because we knowwhat is sent over the wire, not serializing everything

3. Reusability: If there are reusabletypes, we have the option of segregating them into a different file.

4. Versioning: If we leave WSDLto be generated, then we seem to lose control over the versioning multipleWSDLs for the same web service.

5. Can you describe Inversion ofcontrol / dependency injection?

Well.. When Ihear Dependency Injection, I think of Martin Fowler. It is a separatething that I am one of his big fans. I simply love his talks, his articlesthough lot of them are tangential to me J

Sorry.. Backto the point.

Before Idescribe this, I would like to raise the difference between the library and aframework. If we consider library as a dull way of simply using things in aprocedural way, there comes the framework which means taking the control overand maintaining the flow and in a way becoming the heart of the application.Once we realise this, it is relatively easy to explain inversion of control.

As it says,inversion of control means the traditional flow of control is inverted andexpects that the module or the framework actually makes the call into theapplication specific libraries. Some of the ways to actually get this doneinclude publisher & subscriber mechanism and the dependency Injection.

Without goinginto the details of dependency injection, it is easy to see this in factorypattern where the code is implemented against the interface. And only at runtime, we have the actual object injected into the class.

Because the dependencybetween the components is injected at runtime, this is called dependencyinjection. We have two ways to do this. One way is to inject it through theconstructor and the other way to do it is through a setter method.

While thereare multiple advantages in terms of flexibility, there is increased complexityand the loss of code readability. And at the same time, it might be tricky toactually write and run unit tests which would test the dependency injection.

6. Can you describe a 2 patterns anddescribe where and how you would use them?

Out of the ones that are left after I completed the listof questions... My favourites are

1. Observer Pattern

a. This is used to observe the state of the publisher by the set ofsubscribers.

b. We basicallyhave a “Subject and Concrete Subject” and “Observer and ConcreteObserver”classes. Observers subscribes to the events from the subject.

c. Subjectkeeps raising the events whenever there is a state change in the Subject.

d. This is oneof the most commonly used patterns of all. We use this regularly while we workwith ASP.NET applications. We hookonto the events using this model.

e. Along with this,we can actually have the model pinging the views back in MVC pattern using thispattern.

2. FacadePattern

a. Thisprovides simple interface for huge number of complex systems behind the scenes.

b. This is pretty easy to implement. Main class objects are composedinside the facade pattern and whenever the client makes the calls, then itmerely uses the composed objects to get the functionality done.

c. We havecreated quite a lot of facade classes over Microsoft Enterprise Libraries tosimplify the access to the client application code.

7. Can you describe what genericsare in context of .net 2?

Coming fromthe C++ background, I am very much familiar with generics. Once we startworking with Generics, we might soon feel how boring the life would be withoutgenerics.

This conceptallows us to be able to let the class not bind to a particular type and thesame time providing the excellent type safety. Excellent example is the genericcollection in .NET 2.0. This is one feature which has been exploited in .NET3.5 while we work with LINQ.

8. Can you describe error handling in.net?

.NET provides multiple things to be able to deal witherrors in any of the .NET languages.

1. .NETCLR provides the exception handling mechanisms. This has a service called SEH(Structured Exceptional Handling) service that helps us to recover from theexceptions that are thrown by runtime.

a. Typically, it is like try, any number of catch blocks and an optional finallyblock. We have facilities to maintain the stack trace even after we catch theexceptions (by rethrowing the exceptions)

b. We have thefacilities to create the custom exceptions that could be thrown outside.

2. .NETprovides facilities to be able debug, trace information without writing bigamount of code

9. Can you describe anonymous methods?

Anonymousmethods is a way of defining a method body without actually giving a name forit. Basically, cases like a small sized methods and to pass as a parameter to adelegate, we use anonymous methods.

I do notbelieve that CLR is doing anything special for the anonymous methods. It mightthe case where CLR is not even aware of it. Personally, I like to name each ofmy methods and do not want to increase the complexity of my code going foranonymous methods.

However, thisis getting into regular use now c# 3.5 which can potentially involve heavy useof anonymous methods.

10. Whatdoes the term immutable mean?

By definition,immutable object means the state cannot be changed after it has beeninitialized once. In .NET, one of the common examples that we have is thestring object. Under normal execution circumstances and under the executioncontext of .NET, .NET does guarantee that the object value once initializedwould never change in its life time for this string object.

However, as wetry and start creating our own immutable objects - it could be linked lists,trees etc or even a data holder then we might need to follow rigorousapproaches to make sure that what we are creating are indeed immutable objects.Some of the tricks include creating only getter properties and not allow anysetters. And try and make the class sealed as the opportunity arises. Possibly,declare all the private members as private readonly.

11. What is a mock and why do we use them?

Well.. Withoutmocking, we simply cannot do TDD. It is natural for any of the Unit Code tohave dependencies with the external entities like Database, Web services etc.If our unit tests actually get the other dependencies into act, for multiplereasons the TDD would not be possible.

For thisreason, we need to provide the custom implementations (Typically frameworksprovide) that would similar/mock the actual behaviour of the dependencies.Hence, it allows the Unit test to run and at the same time, Unit test is nomore depending on the external forces.

Sorry forgotto answer what is Mock. Formally, mock objects are the custom objects thatwould simulate the behaviour of the real objects in a very controlled way. Andmock objects differ from Fake objects in a way that mock objects could actuallyvalidate the input data and raise the assertions.

12. What is the builder pattern and why would we use it?

Builderpattern is part of the creational patterns where we also have factory,singleton etc. Formally we define builder pattern as “Separate the constructionof a complex object from its representation so that the same constructionprocess can be reused for creating different representations.”. It gets all thebenefits of going for a creational pattern and more as described in thedefinition.

13. What is contract first development?

When creatingweb services, there are two different styles of developing the services.

· Contractfirst:C++ days, WCF is nearly there

· Codefirst:VB days, ASMX implementation

When we are doing contract first, we startwith the WSDL contract first and then code for that contract. This helpsin solving lot of problems including the below.

1. Fragility: No guarantee thatcontract stays the same over the long time

2. Performance is high because we knowwhat is sent over the wire, not serializing everything

3. Reusability: If there are reusabletypes, we have the option of segregating them into a different file.

4. Versioning:If we leave WSDL to be generated, then we seem to lose control overthe versioning multiple WSDLs for the same web service.

14. What is the MVC Pattern; please describe the M, V and C andtheir roles within the pattern?

MVC stands forthe Model View Controller pattern. M – Model, V – View, C – Controller

I was involvedin one of the architectural study that we made for comparing multiple UIarchitectural patterns to come up a mortgage designer UI application. Well, we can discuss that when I am there.

Meanwhile, MVCis an architectural pattern which well renowned and which has maximum number ofinterpretations in the software industry. My understanding says that thecontroller is responsible for the flow between the pages and in most of theimplementations it hosts the presentation logic as well. And the model istypically equivalent to the responsible person to work with the data. Whereas,view is responsible just for showing the content to the user.

In a wellimplemented model, we can see view refresh updates coming from model as wellalong with View himself pulling the data from Model. Controller actuallyco-ordinates the effort starting with listening to the events till the viewgets refreshed.

15. What the SOC principal and its application within codearchitecture?

Frankly, I am not aware of any concept withthe name SOC principal. I will assume it is SOA and explain what I know here.

SOA stands for Service OrientedArchitecture. This is more of an enterprise architectural pattern wherethe particular services are exposed and maintained at an enterprise level.There are set of rules/principles that this architecture is based on. Some ofthem include versioning, reuse.

This architecture encourages loose couplingbetween the components. And at the same time allows for seemless integrationand communication between the components.

16. List and describe some code smells?

Code smell basically is an indicator that something mightbe going wrong. As Martin Fowler says, for purists they will not want to seeany code smells, where as pragmatists take it case by case.

Some of the code smells that I am aware of are...

1. Hugemethods & Huge classes or class with too less code

2. Duplication of code

3. Unusedcode

4. Repeating the switch statements

5. Emptycatch clauses or catch all clauses.

17. What is Inheritance?

In our OOworld, this is the first class citizen. Formally, inheritance is a concept ofspecialization or is-A relationship. Through inheritance we can derive anotherclass from the base class and hence would be able to use the attributes and thebehaviour of the base class. By this, it supports code-reuse.

However, offlate because of issues like long inheritance hierarchies, allowing only singleinheritance, being static in determining the class (compile time binding), Thiscame under lot of fire which led to the design principle of “Favor composition overinheritance”.

This is thefundamental framework behind the features like Overriding.

18. Describe how we might express an expectation in the contextof a N/JUnit Test?

For any unittesting framework, let it be NUnit/Visual Studio Unit testing framework, wehave the assert class that would poses the static methods to evaluate anddescribe the expectations.

We havedifferent type of assetions in NUnit which includes equality asserts,comparision asserts and utility methods like fail and ignore. Apart from these,we have the StringAssert class that provides additional assertions on strings.

19. Describe either the Inversion of Control Principal or FluentInterface?

I haveexplained Inversion of Control Principal earlier. I will explain fluentinterface below.

A fluentinterface is a more fluent way of expressing something.

Example:

MailMessage msg = new MailMessage("srihari","SomeSubject", "Some Body");
SMTPServer server = new SMTPServer("mysmtpserver");
server.Send(msg);

This isn’t afluent interface. Whereas the following one is a fluent interface.

Mail.WithRecipient("srihari")
.WithSubject("SomeSubject")
.WithBody("somebody")
.WithSMTPServer("mysmtpserver")
.Send();

This isimplemented by a central concept of implementing the “With setter methods” andin each of them returning the interface that would point to the object itself.

How to recognize if the email is Spam?

noreply@blogger.com (Srihari Kothapalli) — Fri, 15 Aug 2008 12:24:00 +0000

I was thinking how does outlook know if the email is spam?

Apperantly, it is a strightforward thing. Have a look below.

The most common way of detecting spam is rejecting mail that comes from mail servers known (or believed) to send spam. This is done by taking the IP address of the remote mail server, converting it to a domain name using the ip4r format (a.b.c.d becomes d.c.b.a.lookupzone.com), and doing a lookup to check if that IP address is listed in one of the databases.

http://whatismyipaddress.com/staticpages/index.php/is-my-ip-address-blacklisted

So, the best way to spam is to send the emails with gmail or yahoo accounts :)

Srihari

How to limit the data precision to 2 in SQL Server

noreply@blogger.com (Srihari Kothapalli) — Thu, 14 Aug 2008 14:16:00 +0000

I assume there must be a quicker way to do this. If not, following will do the trick.

let us say we want to have the output for 2012 as 20.12, we can use the following query.

select cast((cast(2012 as numeric(10,2))/100) as numeric(10,2)) as Amount

My brother celebrating Indepence day

noreply@blogger.com (Srihari Kothapalli) — Thu, 14 Aug 2008 12:03:00 +0000

Tomorrow is INDIAns Independence day !!! What a great feeling...

My brother celebrating independence day pics captured and posted at the link below.

http://picasaweb.google.com/SndSuresh/SureshNewPhotos

How to do sequential ranking for each item in group in SQL server

noreply@blogger.com (Srihari Kothapalli) — Wed, 13 Aug 2008 14:34:00 +0000

Problem:

I have a table with the name Employee. This has columns Person Name, Date Of Birth. Now I want to rank all the persons in each month ordered by person name.

create table People
(
[name] varchar(20),
dob datetime
)

Assume that we have the input data as ...

insert into people values('Srihari', '10-20-2008');
insert into people values('Suresh', '10-21-2008');
insert into people values('Kavitha', '10-19-2008');
insert into people values('Srinu', '12-18-2008');

Srihari - 20/10/2008
Suresh - 21/10/2008
Kavitha - 19/10/2008
Brahmaiah - 19/12/2008
Srinu - 18/12/2008

I want to have the following output.

Kavitha - 10 - 1
Srihari - 10 - 2
Suresh - 10 - 3
Srinu - 12 - 1
Brahmaiah - 12 - 2

As you can see none of the SQL Server rank functions can be used straight away to achieve this functionality.

If we use rank...

select *, rank() over (order by dob)
from people

we get the following output..

Srihari Oct 20 2008 1
Suresh Oct 21 2008 1
Kavitha Oct 19 2008 1
Brahmaiah Dec 19 2008 4
Srinu Dec 18 2008 4

This is obviously not what wanted...

Let us try with some thing else now.. Row_Number???

select *, row_number() over (order by dob)
from people

Kavitha Oct 19 2008 1
Srihari Oct 20 2008 2
Suresh Oct 21 2008 3
Srinu Dec 18 2008 4
Brahmaiah Dec 19 2008 5

How do we solve our original problem?

Well... What we can do is to mix both of these to acheive the required functionality.

Try the below one..

select *, (row_number() over (order by dob) -rank() over (order by datepart(month,dob)) + 1) as GroupRank
from people

Kavitha Oct 19 2008 1
Srihari Oct 20 2008 2
Suresh Oct 21 2008 3
Srinu Dec 18 2008 1
Brahmaiah Dec 19 2008 2

Then, we can actually rank them after ordering individual group. Let me know if you can not go further to do that. I am happy to help you.

Ranking functions in SQL Server

noreply@blogger.com (Srihari Kothapalli) — Wed, 13 Aug 2008 11:57:00 +0000

I found a real good example to understand different ranking functions in SQL Server. I thought of sharing it with every body.

USE AdventureWorks; GO SELECT c.FirstName, c.LastName ,ROW_NUMBER() OVER (ORDER BY a.PostalCode) AS 'Row Number' ,RANK() OVER (ORDER BY a.PostalCode) AS 'Rank' ,DENSE_RANK() OVER (ORDER BY a.PostalCode) AS 'Dense Rank' ,NTILE(4) OVER (ORDER BY a.PostalCode) AS 'Quartile' ,s.SalesYTD, a.PostalCode FROM Sales.SalesPerson s INNER JOIN Person.Contact c ON s.SalesPersonID = c.ContactID INNER JOIN Person.Address a ON a.AddressID = c.ContactID WHERE TerritoryID IS NOT NULL AND SalesYTD <> 0;

FirstName LastName Row Number Rank Dense Rank Quartile SalesYTD PostalCode
Maciej Dusza 1 1 1 1 4557045 98027
Shelley Dyck 2 1 1 1 5200475 98027
Linda Ecoffey 3 1 1 1 3857164 98027
Carla Eldridge 4 1 1 1 1764939 98027
Carol Elliott 5 1 1 2 2811013 98027
Jauna Elson 6 6 2 2 3018725 98055
Michael Emanuel 7 6 2 2 3189356 98055
Terry Eminhizer 8 6 2 3 3587378 98055
Gail Erickson 9 6 2 3 5015682 98055
Mark Erickson 10 6 2 3 3827950 98055
Martha Espinoza 11 6 2 4 1931620 98055
Janeth Esteves 12 6 2 4 2241204 98055
Twanna Evans 13 6 2 4 1758386 98055

http://msdn.microsoft.com/en-gb/library/ms189798.aspx

How to get the Logged in Machine Name from the user name

noreply@blogger.com (Srihari Kothapalli) — Mon, 11 Aug 2008 19:56:00 +0000

The steps would help us to get the task done.

1) Search the AD and locate the user by their username
2) Determine the user's home drive location
3) Split that string down to get the servername out of the UNC path
4) Query the sessions on that server to see if the user is currently connected (any logged on user should be showing an open connection to their home drive server)
5) Use that info to determine the PC name/IP that they are connecting from

Refer: http://forums.whirlpool.net.au/forum-replies-archive.cfm/549971.html for more details.

Why hardcoding connection strings is bad

noreply@blogger.com (Srihari Kothapalli) — Thu, 07 Aug 2008 09:42:00 +0000

Today, i was chatting to one of my friends who is just starting up his career. I had to answer to his question on why hardcoding the connection strings is bad. I thought I could paste that chat here in my blog to help others who are just starting up.

suresh: you said ealrier avoding hard coding of connection strings .
means
srihari: yes
suresh: should i place it under dbcontsnas.cs file
is that fine ?
srihari: ok let me tell you some thing
understand it to your heart
assume that you have hard coded it in a function like you did
suresh: i can feel it
srihari: let us you have 100 such functions like that to write
suresh: i can understand the problem
srihari: listen
suresh: Ok
srihari: what is the problem
tell me
suresh: if we hard code it
we can not reuse the sme string
if we place it under file
srihari: good
suresh: we can use it among multiple functions where ever it is required
srihari: good
suresh: becaus of that i asked you can i assign this string to some variable of type string and place it under dbconstants.cs file so that we can use that varaible where ever it is necessary
srihari: good point
let me tell you bit more
suresh: ok
srihari: now, we will talk like professionals
suresh: ok
srihari: if some body asks me .. i will tell them the below
List of problems
1. By hardcoding the connection string, your project can not be used against a different database.
assume that you have a live database, test database
suresh: yes
srihari: what will happen if you hardcode test database name into the source code
you can not take that project live :)
is this cleart
suresh: yes
srihari: so, this basically means.. it has to come from a configuration file.. which can be modified ..
this should not be part of the source code
is this clear
suresh: yes
web.config is not there
i thought in that way also
as it is class library project
srihari: thats why there is some thing called app.config
did you check that
suresh: ok
no
srihari: also, remember one thing
you are going to ultimately use this from a web site right which will have web.config
suresh: yes
srihari: those configuration parameters will take effect
suresh: ok
srihari: Now, let us talk about the second problem
suresh: ok
srihari: 2. Let us say you have hardcoded the connection string, if the database username and password change
your application will stop working
suresh: Yes
srihari: then we will have to rebuilt the problem
sorry code
suresh: ok
unerstood
srihari: basically, we have to make sure that code should never be changed
because for any code change that we make to the code, we have to go through the entire cycle of testing, deployment.
this is very costly
suresh: ok
srihari: so, when we design... we have to code so that ... there will be very minimal number of changes
Is this clear
suresh: Yes
srihari: good
suresh: it is clear
srihari: now let me tell you a principle
this priciple is a GURU in OOPS design
This is named as "Open Closed Principle"
suresh: ooh
i heared about this already
akkayya answered for this
srihari: This means.. "your code should always be OPEN for modifications and CLOSED for the changes"
so by hard coding our connections string, we are violating the open closed principle
because our code is not closed for changes
sorry bujji
my mistkae
"your code should always be OPEN for EXTENSION and CLOSED for the changes"
extension.. modification kadu
suresh: ok
srihari: there is lot more to learn in this
but for now.. understand that we are violating this principle.. because our code is not closed for changes
changes include like datbase connections
is this clear
suresh: our code is not closed for changes
srihari: yes
suresh: i did not understand this
our code is closed for changes
srihari: closed ante... if we need to change the connections string.. we need to change the code right
suresh: i did not understand this
yes
srihari: so, our code is not closed against the database connection string modification
if we have to modify the database connection string, we need to go inside the code and change it
as i told you earlier, changing the code is very costly
is this clear now
ask as many questions as possible
understand this concept to your heart
suresh: so we need to minizes the changes to the possible extent
srihari: yes
good point
then the code is more maintainable
suresh: Ok
srihari: in any software project... 50% cost is in initial development
and 50% cost in maintaience
so, it is very important to think about maintainence while we are development
if we do not follow good design principles.. we will be scerwed up during maintainence
is this clear
suresh: ok
srihari: good
now the third problem
3. If we hard code the connection string inside the function,
there will be multiple places where the same connection string will be present
tomorrow if there is a need to modify the connection string, we need to make sure that all the places are updated correctly
suresh: ye
srihari: this is very very difficult
imagine you have developed 10000 lines of code
and you are looking at the code after 1 year
there is no way, that you will be able to know all the places that connection strings are present
suresh: yes
srihari: then just for changing the connection strng, you will have to spend millions of pounds
this is a maintainence nightmare
is this clear
suresh: maintainence nightmare ?
night mare
srihari: nightmare.. pida kala
suresh: ok
srihari: :)
now.. the fourth problem
4. This is not quite applicable for the connection strings
but in general applicable for the strings that will be shown to the user
Imagine today you have released your applicaiton in english
every thing works fine..
what happens, if you want to change the language to telugu
you need to go into the code and modify all the strings inside the code
suresh: yes
srihari: against this is inconsitent, maintainence problem
is this clear
suresh: yes
srihari: thats why we have got resource files (satillite assemblies)
suresh: yes
srihari: to solve this problem.. what we do is to keep the keys in the constants file
and get the values from the resource dlls
suresh: ok
srihari: in general, fourth point we do not care much
because most of the applications that we write are targetted only for english lands
suresh: yes
srihari: in microsoft, i used to work with resoruce assemblies

Hope this helped.

Windows Authentication followed by Forms Authentication in ASP.NET

noreply@blogger.com (Srihari Kothapalli) — Tue, 05 Aug 2008 15:50:00 +0000

Today, I was search for a way to find out if we can do both windows authentication and forms authentication if the former one fails. The answer is yes.. but it is not stright forward.. It appears that can be achieved only by creating another virtual directory.

Check the below article. http://aspalliance.com/553

Regards,

Srihari

New certifications are available to take on .NET 3.5

noreply@blogger.com (Srihari Kothapalli) — Tue, 05 Aug 2008 11:50:00 +0000

Me and my wife are MCPD certified Enterprise applications Developers. And, we are constantly looking to extend our experience and prove it by taking certifications.

Here are the new certifications in .NET WPF, WCF, WF.

Exam 70-502: TS: Microsoft .NET Framework 3.5, Windows Presentation Foundation Application Development (English)
•
Exam 70-503: TS: Microsoft .NET Framework 3.5, Windows Communication Foundation Application Development (English)
•
Exam 70-504: TS: Microsoft .NET Framework 3.5, Windows Workflow Foundation Application Development (English)

Srihari

Executing a batch file programmatically

noreply@blogger.com (Srihari Kothapalli) — Mon, 04 Aug 2008 15:39:00 +0000

While we want to do any kind of database automation.. as we know that.. C#/.NET does not provide any inbuilt mechanism to be execute any database script.. We might have to rely on SQLCMD.EXE or oSQL if it is 2000.

I wrote the following piece of code which could execute my sqlcmd executable file/batch files through my c# program. I hope this could help you.

private void RunSqlScriptInternal(DatabaseNode database, String SqlToExecute, String OutSqlFileName)
{
Console.WriteLine("Started executing script: " + SqlToExecute);
String parentDir = Directory.GetParent(OutSqlFileName).FullName;
if (!Directory.Exists(parentDir))
{
Directory.CreateDirectory(parentDir);
}
try
{
using (Process proc = new Process())
{
if (!database.integratedAuth)
{
proc.StartInfo.Arguments = @"-U" + database.userName + " -P" + database.password + " -S" + database.serverName + " -d" + database.databaseName + @" -i""" + SqlToExecute + @""" -o""" + OutSqlFileName + @""" -h -1 -W";
}
else
{
proc.StartInfo.Arguments = @"-E -S" + database.serverName + " -d" + database.databaseName + @" -i""" + SqlToExecute + @""" -o""" + OutSqlFileName + @""" -h -1 -W";
}
proc.StartInfo.FileName = @"""C:\Program Files\Microsoft SQL Server\90\Tools\Binn\SqlCmd.exe""";
proc.StartInfo.UseShellExecute = false;
proc.StartInfo.RedirectStandardOutput = true;
proc.StartInfo.RedirectStandardError = true;
proc.StartInfo.CreateNoWindow = true;
proc.Start();
Console.ForegroundColor = ConsoleColor.DarkBlue;
String output = proc.StandardOutput.ReadToEnd();
if (!String.IsNullOrEmpty(output)) Console.WriteLine(output);
proc.WaitForExit();
output = proc.StandardOutput.ReadToEnd();
if (!String.IsNullOrEmpty(output)) Console.WriteLine(output);
Console.ForegroundColor = ConsoleColor.Red;
output = proc.StandardError.ReadToEnd();
if (!String.IsNullOrEmpty(output)) Console.WriteLine(output);
Console.ForegroundColor = ConsoleColor.White;
proc.Close();
}
}
catch (Exception e)
{
Console.ForegroundColor = ConsoleColor.Red;
Console.WriteLine(e.Message);
Console.ForegroundColor = ConsoleColor.White;
}
Console.WriteLine("Finished executing script. ");
Console.WriteLine();
}

Comparing database schema and data between two databases

noreply@blogger.com (Srihari Kothapalli) — Mon, 04 Aug 2008 15:34:00 +0000

Recently, I had come across this task of comparing data between two databases. I thought, I could share the research through this blog.

Options that we have in hand:

1. Link the two databases together, then it is strightforward.
2. Otherwise, you can use the visual studio data professional product to compare the data betwen two databases. I am not sure if they are exposing any APIs to do this. I would guess no.
And, the third option is to pull the data into a local database and compare the tables.

Piece of code, that i have used to compare two tables.

declare @SQL varchar(8000)
declare @T1ColumnList varchar(1000)
declare @t2ColumnList varchar(1000)
declare @table1 varchar(100)
declare @table2 varchar(100)
set @T1ColumnList='$COLUMN_NAMES$'
set @table1='$TABLE1$'
set @table2='$TABLE2$'
SET @T2ColumnList = @T1ColumnList
set @SQL = 'SELECT ''' + @table1 + ''' AS TableName, ' + @t1ColumnList + ' FROM ' + @Table1 + ' UNION ALL SELECT ''' + @table2 + ''' As TableName, ' + @t2ColumnList + ' FROM ' + @Table2
set @SQL = 'SELECT Max(TableName) as TableName, ' + @t1ColumnList + ' FROM (' + @SQL + ') A GROUP BY ' + @t1ColumnList + ' HAVING COUNT(*) = 1'
SET NOCOUNT ON
exec (@SQL)

Hope you found this useful.

Agile Estimation and Planning

noreply@blogger.com (Srihari Kothapalli) — Sun, 13 Jul 2008 17:18:00 +0000

I found the following really good videos that helps you understand how estimation inside Agile way of development works.

1. http://www.youtube.com/watch?v=fb9Rzyi8b90

2. http://www.youtube.com/watch?v=jeT0pOVg0EI

Srihari

Foundations of Better Programming

noreply@blogger.com (Srihari Kothapalli) — Fri, 04 Jul 2008 20:45:00 +0000

I came across this book which I found very useful. This covers excellent topics like Agile programming, unit testing, mock objects, memory management, object relational mapping, and more. I thought this could help while we design applications especially in agile development.

http://codebetter.com/files/folders/codebetter_downloads/entry179694.aspx

Trip to Windsor Castle

noreply@blogger.com (Srihari Kothapalli) — Mon, 30 Jun 2008 11:52:00 +0000

Please find some snaps at the below link.

http://picasaweb.google.com/srihari.kothapalli/WindsorCastle

Srihari

when to use linq to sql, entity framework

noreply@blogger.com (Srihari Kothapalli) — Sun, 29 Jun 2008 23:25:00 +0000

http://msdn.microsoft.com/en-us/library/cc161164.aspx

Helpful links to learn ASP.NET Session Management internals

noreply@blogger.com (Srihari Kothapalli) — Sat, 28 Jun 2008 14:54:00 +0000

Useful set of links to get started with understanding sessions from basic to in depth explanations

Session Management basics

http://www.codeproject.com/KB/aspnet/SessionManagementAspNet.aspx

Underpinnings of the Session State Implementation in ASP.NEThttp://msdn.microsoft.com/en-us/library/aa479041.aspx

Session Internals: http://www.codeproject.com/KB/session/ASPNETSessionInternals.aspx
A Session Data Management Tool
Introducing a .NET web service to manage session data for multiple applications
The Defibrillator: Keeping ASP.NET Session Alive Ad Infinitum
With only two lines of code (no joking) keep your session alive as long as the user stays on your site.
Prevent Session Timeout in ASP.NET
Simple code to prevent a sesison from timing out while a user enters data or edits HTML etc.
Offering a better (ODP Compliant) ASP.NET Session object
This article discusses the problems and the available solutions for maintenance and utilization of ASP.NET session state of the .NET session object.
Using Attributes for encapsulating ASP.NET Session and ViewState variables
A generic and type-safe way to encapsulate ASP.NET Session and ViewState variables.

Is union supported in C#?

noreply@blogger.com (Srihari Kothapalli) — Sun, 22 Jun 2008 16:50:00 +0000

Today, one of my friends asked me a tricky question. Is union supported in c#?
Answer is No. But we can actually simulate it with a concept called StructLayout.
To give a little background, in .NET the way all the fields are positions while allocating the memory is not in our control by default. (Please note this applies to classes only. For structs, it is actually the way we declare the fields.)

There are three commands that CLR can obey in terms of the field layout.

1. Layout.Sequential (Default for Struct)
2. Layout.Auto (Default for Class)
3. Layout.Explicit (We can use this to create unions)

I guess, all the above three are self explanatory. We can specify this to CLR using a class attribute called StructLayoutAttribute.

[StructLayout(LayoutKind.Sequential, Size=8)]class myClass
{
}

We have one more field attribute called “FieldOffset” through which we can specify the relative address location for the field.

Given this information, look at the following.

[StructLayout(LayoutKind.Explicit, Size=8)]
struct UValue
{
[FieldOffset(0)] public char _cval;
[FieldOffset(0)] public int _ival;
[FieldOffset(0)] public double _dval;
}

Can you sense a Union here? Yes. You are right. This is precisely the union that we had in C and C++. We actually have much more flexibility.

This actually is equivalent to the following union in C.

union UValue
{
char _cval;
int _ival;
double _dval;
};

Hope this helped you.

Useful tip for guys starting to use WCF

noreply@blogger.com (Srihari Kothapalli) — Sat, 21 Jun 2008 22:51:00 +0000

When you are creating your WCF files using the class library and subsequently deploy it in the IIS, please note that by default project configuration says “the dll would be created under the bin\debug”. This is a problem for the IIS. IIS only looks for the dlls under bin or the compiles the code under app_code directory. So, inorder to solve this problem, either you can manually copy the dll generated to bin folder or actually change the settings in the project properties to generate the dll to bin directory.

Tip to open docx files with msword 2003

noreply@blogger.com (Srihari Kothapalli) — Sat, 21 Jun 2008 16:16:00 +0000

Kavitha has recently blogged out a hack through which you can open the docx files with doc software.

Here are the steps.

1. Rename your docx file to zip.
2. Unzip the file.
3. Go to document.xml and open this file with microsoft word 2003 that you have.
4. You will then see the file, with all the formatting missed but still not readable.

You can directly go to that blog.

Srihari