Skip to content
  • Categories
  • Recent
  • Tags
  • Popular
  • World
  • Users
  • Groups
Skins
  • Light
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • Dark
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • Default (No Skin)
  • No Skin
Collapse
Code Project
CODE PROJECT For Those Who Code
  • Home
  • Articles
  • FAQ
Community
  1. Home
  2. General Programming
  3. C#
  4. The Switch Statement

The Switch Statement

Scheduled Pinned Locked Moved C#
csscomtoolsarchitecturequestion
29 Posts 14 Posters 2 Views 1 Watching
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • J Jammer 0

    Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

    Jammer My Blog | Article(s)

    P Offline
    P Offline
    PIEBALDconsult
    wrote on last edited by
    #9

    Yeah, what they said. Or, start using Dictionary<sometype,delegate> to simulate switches: http://www.codeproject.com/KB/tips/DictorionaryEnumDelegate.aspx[^]

    G 1 Reply Last reply
    0
    • J Jammer 0

      Heh heh ... to be fair, I've been learning C# since February and it wouldn't be the first time that I've come across something in the language and other more experienced folks have said to me "oooh ... you really don't want to use that ... that's bad" or similar. And like everything there is more than one way to skin a cat. I'm sure I could design out the need for the switch statements I've used so far if I really wanted too ...

      Jammer My Blog | Article(s)

      M Offline
      M Offline
      Mark Salsbery
      wrote on last edited by
      #10

      IMO it comes down to the resulting compiled code. The ability of a computer to branch based on a decision is powerful, and taken for granted these days. So, when designing a high level language, where you want the ability to branch based on a list of comparisons, what are the alternatives? One can do it explicitly with if/then/else but that can be inefficient for the decision at the end of a big list. To avoid that inefficiency some kind of look-up table may be more efficient. Theoretically, using the switch semantics, a good compiler could recognize a big list and produce more efficient branching code - possibly using a look-up table of some kind. The restrictions of the switch statement allows this. Here's a C++ example ( completely OT for this forum, but demonstrates my point):

      int i = 6;
      
      if (i == 0)
      	TRACE0("0");
      else if (i == 1)
      	TRACE0("1");
      else if (i == 2)
      	TRACE0("2");
      else if (i == 3)
      	TRACE0("3");
      else if (i == 4)
      	TRACE0("4");
      else if (i == 5)
      	TRACE0("5");
      else if (i == 6)
      	TRACE0("6");
      else 
      	TRACE0("n");
      
      switch (i)
      {
      case 0:
      	TRACE0("0");
      	break;
      case 1:
      	TRACE0("1");
      	break;
      case 2:
      	TRACE0("2");
      	break;
      case 3:
      	TRACE0("3");
      	break;
      case 4:
      	TRACE0("4");
      	break;
      case 5:
      	TRACE0("5");
      	break;
      case 6:
      	TRACE0("6");
      	break;
      default:
      	TRACE0("n");
      	break;
      }
      

      The resulting if/else machine code does six comparisons to get to the resulting branch. The resulting switch machine code does one comparison to see if i is > 6, and if it isn't, branches directly to the correct case using only TWO machine code instructions (via a look-up table)

      0041D9F5 mov ecx,dword ptr [ebp-0AE4h]
      0041D9FB jmp dword ptr (41F8CCh)[ecx*4]

      switch may be ugly in source code, but the compiler can certainly take advantage of it to produce efficient code. And I don't buy the "purist programmer" arguments...what purist would use C# (or any other higher level language)? ;) It still all comes down to knowing the language you're using and choosing the right instructions for the given situation... Mark

      Mark Salsbery Microsoft MVP - Visual C++ :java:

      J 1 Reply Last reply
      0
      • M Mark Salsbery

        IMO it comes down to the resulting compiled code. The ability of a computer to branch based on a decision is powerful, and taken for granted these days. So, when designing a high level language, where you want the ability to branch based on a list of comparisons, what are the alternatives? One can do it explicitly with if/then/else but that can be inefficient for the decision at the end of a big list. To avoid that inefficiency some kind of look-up table may be more efficient. Theoretically, using the switch semantics, a good compiler could recognize a big list and produce more efficient branching code - possibly using a look-up table of some kind. The restrictions of the switch statement allows this. Here's a C++ example ( completely OT for this forum, but demonstrates my point):

        int i = 6;
        
        if (i == 0)
        	TRACE0("0");
        else if (i == 1)
        	TRACE0("1");
        else if (i == 2)
        	TRACE0("2");
        else if (i == 3)
        	TRACE0("3");
        else if (i == 4)
        	TRACE0("4");
        else if (i == 5)
        	TRACE0("5");
        else if (i == 6)
        	TRACE0("6");
        else 
        	TRACE0("n");
        
        switch (i)
        {
        case 0:
        	TRACE0("0");
        	break;
        case 1:
        	TRACE0("1");
        	break;
        case 2:
        	TRACE0("2");
        	break;
        case 3:
        	TRACE0("3");
        	break;
        case 4:
        	TRACE0("4");
        	break;
        case 5:
        	TRACE0("5");
        	break;
        case 6:
        	TRACE0("6");
        	break;
        default:
        	TRACE0("n");
        	break;
        }
        

        The resulting if/else machine code does six comparisons to get to the resulting branch. The resulting switch machine code does one comparison to see if i is > 6, and if it isn't, branches directly to the correct case using only TWO machine code instructions (via a look-up table)

        0041D9F5 mov ecx,dword ptr [ebp-0AE4h]
        0041D9FB jmp dword ptr (41F8CCh)[ecx*4]

        switch may be ugly in source code, but the compiler can certainly take advantage of it to produce efficient code. And I don't buy the "purist programmer" arguments...what purist would use C# (or any other higher level language)? ;) It still all comes down to knowing the language you're using and choosing the right instructions for the given situation... Mark

        Mark Salsbery Microsoft MVP - Visual C++ :java:

        J Offline
        J Offline
        Jammer 0
        wrote on last edited by
        #11

        Great example! Thanks for this ...

        Jammer My Blog | Article(s)

        M 1 Reply Last reply
        0
        • J Jammer 0

          Great example! Thanks for this ...

          Jammer My Blog | Article(s)

          M Offline
          M Offline
          Mark Salsbery
          wrote on last edited by
          #12

          Heh no problem. I work alone at home - I don't get to have these water-cooler discussions with colleagues :) Cheers!

          Mark Salsbery Microsoft MVP - Visual C++ :java:

          J 1 Reply Last reply
          0
          • M Mark Salsbery

            Heh no problem. I work alone at home - I don't get to have these water-cooler discussions with colleagues :) Cheers!

            Mark Salsbery Microsoft MVP - Visual C++ :java:

            J Offline
            J Offline
            Jammer 0
            wrote on last edited by
            #13

            heh ... they can be really beneficial! Although I'm still at a stage in learning .NET where I come away with more questions than answers! Doh!

            Jammer My Blog | Article(s)

            1 Reply Last reply
            0
            • J Jammer 0

              Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

              Jammer My Blog | Article(s)

              C Offline
              C Offline
              carbon_golem
              wrote on last edited by
              #14

              Not to beat the "dead horse" here, but I didn't see anyone bring up the induced complexity that the switch adds. See complexity definition here.[^] The higher the complexity score, the more trouble maintenance and testing become. The switch is very efficient at adding additional paths in the code (cases - 1). While yes, the readability is there, one could argue at the same time a certain amount of flexibility is lost at the expense. There are more or less feasible ways around using a switch or bank of if/else altogether - using the chain-of-responsibility pattern has worked for me in some situations. And certainly it has made unit testing easier. I might correct your colleague's statement and submit that maybe the architecture isn't "wrong" by using a switch, it just might be pointing out that there are higher quality alternatives. Scott P.

              "Simplicity carried to the extreme becomes elegance."
              -Jon Franklin

              L realJSOPR 2 Replies Last reply
              0
              • J Jammer 0

                Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

                Jammer My Blog | Article(s)

                L Offline
                L Offline
                led mike
                wrote on last edited by
                #15

                Jammer wrote:

                What do other folks think of this?

                It's not Object Oriented and the following is a section of my own notes to myself (quotes ) from Kent Becks book "Implementation Patterns"Conditional (page 36)

                If/then and switch statements are the simplest form of
                instance –specific behavior….

                The more paths through a program the less likely the program
                is to be correct….The proliferation of conditionals reduces reliability….This
                problem is compounded when conditionals are duplicated.

                These problems can all be eliminated by converting the
                conditional logic to messages, either with subclasses or delegation.


                Last modified: 5hrs 28mins after originally posted --

                led mike

                M 1 Reply Last reply
                0
                • C carbon_golem

                  Not to beat the "dead horse" here, but I didn't see anyone bring up the induced complexity that the switch adds. See complexity definition here.[^] The higher the complexity score, the more trouble maintenance and testing become. The switch is very efficient at adding additional paths in the code (cases - 1). While yes, the readability is there, one could argue at the same time a certain amount of flexibility is lost at the expense. There are more or less feasible ways around using a switch or bank of if/else altogether - using the chain-of-responsibility pattern has worked for me in some situations. And certainly it has made unit testing easier. I might correct your colleague's statement and submit that maybe the architecture isn't "wrong" by using a switch, it just might be pointing out that there are higher quality alternatives. Scott P.

                  "Simplicity carried to the extreme becomes elegance."
                  -Jon Franklin

                  L Offline
                  L Offline
                  led mike
                  wrote on last edited by
                  #16

                  carbon_golem wrote:

                  there are higher quality alternatives.

                  Agreed. See my post below, that you just beat to the wire! Damn you! :laugh:

                  led mike

                  1 Reply Last reply
                  0
                  • J Jammer 0

                    Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

                    Jammer My Blog | Article(s)

                    realJSOPR Offline
                    realJSOPR Offline
                    realJSOP
                    wrote on last edited by
                    #17

                    Well, your colleague is a moron, and it sound s like his father should have used a switch on him more often when he was a child.

                    "Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
                    -----
                    "...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001

                    J 1 Reply Last reply
                    0
                    • C carbon_golem

                      Not to beat the "dead horse" here, but I didn't see anyone bring up the induced complexity that the switch adds. See complexity definition here.[^] The higher the complexity score, the more trouble maintenance and testing become. The switch is very efficient at adding additional paths in the code (cases - 1). While yes, the readability is there, one could argue at the same time a certain amount of flexibility is lost at the expense. There are more or less feasible ways around using a switch or bank of if/else altogether - using the chain-of-responsibility pattern has worked for me in some situations. And certainly it has made unit testing easier. I might correct your colleague's statement and submit that maybe the architecture isn't "wrong" by using a switch, it just might be pointing out that there are higher quality alternatives. Scott P.

                      "Simplicity carried to the extreme becomes elegance."
                      -Jon Franklin

                      realJSOPR Offline
                      realJSOPR Offline
                      realJSOP
                      wrote on last edited by
                      #18

                      carbon_golem wrote:

                      it just might be pointing out that there are higher quality alternatives.

                      You mean like "goto"? LOLOL

                      "Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
                      -----
                      "...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001

                      1 Reply Last reply
                      0
                      • realJSOPR realJSOP

                        Well, your colleague is a moron, and it sound s like his father should have used a switch on him more often when he was a child.

                        "Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
                        -----
                        "...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001

                        J Offline
                        J Offline
                        Jammer 0
                        wrote on last edited by
                        #19

                        Whoa ... it wasn't my colleague that stated this ... he was passing on things he'd read/heard ... it was a light-hearted conversation! He is far from a Moron as well, a very good developer in fact.

                        Jammer My Blog | Article(s)

                        1 Reply Last reply
                        0
                        • J Jammer 0

                          Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

                          Jammer My Blog | Article(s)

                          G Offline
                          G Offline
                          Guffa
                          wrote on last edited by
                          #20

                          Jammer wrote:

                          He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere.

                          That probably comes from someone who don't know how to use a switch properly... What seems more elegant? This:

                          int result;
                          SomeClass.SomeEnum value = SomeClass.GetValue();
                          if (value == SomeClass.SomeEnum.Option1 || value == SomeClass.SomeEnum.Option2) {
                          result = 1;
                          } else if (value == SomeClass.SomeEnum.Option3) {
                          result = 2;
                          } else if (value == SomeClass.SomeEnum.Option4 || value == SomeClass.SomeEnum.Option5 || value == SomeClass.SomeEnum.Option6) {
                          result = 3;
                          } else {
                          result = 4;
                          }

                          or this:

                          int result;
                          switch (SomeClass.GetValue()) {
                          case SomeClass.SomeEnum.Option1:
                          case SomeClass.SomeEnum.Option2: result = 1; break;
                          case SomeClass.SomeEnum.Option3: result = 2; break;
                          case SomeClass.SomeEnum.Option4:
                          case SomeClass.SomeEnum.Option5:
                          case SomeClass.SomeEnum.Option6: result = 3; break;
                          default: result = 4; break;
                          }

                          Besides, the switch solution is faster.

                          Despite everything, the person most likely to be fooling you next is yourself.

                          1 Reply Last reply
                          0
                          • P PIEBALDconsult

                            Yeah, what they said. Or, start using Dictionary<sometype,delegate> to simulate switches: http://www.codeproject.com/KB/tips/DictorionaryEnumDelegate.aspx[^]

                            G Offline
                            G Offline
                            Guffa
                            wrote on last edited by
                            #21

                            Actually, if you have more than just a few case labels in a switch (four IIRC), it's implemented using a dictionary.

                            Despite everything, the person most likely to be fooling you next is yourself.

                            P 1 Reply Last reply
                            0
                            • L led mike

                              Jammer wrote:

                              What do other folks think of this?

                              It's not Object Oriented and the following is a section of my own notes to myself (quotes ) from Kent Becks book "Implementation Patterns"Conditional (page 36)

                              If/then and switch statements are the simplest form of
                              instance –specific behavior….

                              The more paths through a program the less likely the program
                              is to be correct….The proliferation of conditionals reduces reliability….This
                              problem is compounded when conditionals are duplicated.

                              These problems can all be eliminated by converting the
                              conditional logic to messages, either with subclasses or delegation.


                              Last modified: 5hrs 28mins after originally posted --

                              led mike

                              M Offline
                              M Offline
                              Mark Salsbery
                              wrote on last edited by
                              #22

                              :omg:

                              Mark Salsbery Microsoft MVP - Visual C++ :java:

                              L 1 Reply Last reply
                              0
                              • J Jammer 0

                                Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

                                Jammer My Blog | Article(s)

                                C Offline
                                C Offline
                                Colin Angus Mackay
                                wrote on last edited by
                                #23

                                It depends on the situation. There are definitely situations where I've seen needless use of switch statements, and others that would have benefitted from a switch statement. If you find that you are switching on a specific property of a class a lot within that class then it can be an indication that you might want to refactor the code into a class hierarchy. Each derived class handles a specific case that was in the switch statement.

                                Recent blog posts: *Method hiding Vs. overriding *Microsoft Surface *SQL Server / Visual Studio install order My Blog

                                1 Reply Last reply
                                0
                                • M Mark Salsbery

                                  :omg:

                                  Mark Salsbery Microsoft MVP - Visual C++ :java:

                                  L Offline
                                  L Offline
                                  led mike
                                  wrote on last edited by
                                  #24

                                  Wow what a mess. Fixed. Thanks

                                  1 Reply Last reply
                                  0
                                  • G Guffa

                                    Actually, if you have more than just a few case labels in a switch (four IIRC), it's implemented using a dictionary.

                                    Despite everything, the person most likely to be fooling you next is yourself.

                                    P Offline
                                    P Offline
                                    PIEBALDconsult
                                    wrote on last edited by
                                    #25

                                    Yeah, more than five, as you showed back in July. I keep forgetting that. On the other hand, using your own Dictionary gives you more flexibility and the ability to use it in multiple places (if you can think of such a use).

                                    1 Reply Last reply
                                    0
                                    • J Jammer 0

                                      Hi All, I was recently talking to a colleague at work and the topic came round to using swith(). He said that he'd read/heard more than once that if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture elsewhere. What do other folks think of this? Cheers,

                                      Jammer My Blog | Article(s)

                                      N Offline
                                      N Offline
                                      N a v a n e e t h
                                      wrote on last edited by
                                      #26

                                      As others said, switch statement has many benefits. So use it wisely. You should read something about method cohesion[^].

                                      Navaneeth How to use google | Ask smart questions

                                      1 Reply Last reply
                                      0
                                      • M Mark Salsbery

                                        Jammer wrote:

                                        if you find yourself using a switch you can more or less be assured that its because you've done something wrong in your architecture

                                        That makes so much sense considering it's been around since C and the C# founders decided to include it in C# just for bad software architects to use...yeah... :rolleyes:

                                        Mark Salsbery Microsoft MVP - Visual C++ :java:

                                        V Offline
                                        V Offline
                                        Vengatachalapathy Palanivel
                                        wrote on last edited by
                                        #27

                                        Switch case is better in performance than if-else-if ladder. While using switch case, compiler optimizes the code for the fastest execution. you can check this Url also: http://www.blackwasp.co.uk/SpeedTestIfElseSwitch.aspx[^]

                                        Regards, Vengat P

                                        M 1 Reply Last reply
                                        0
                                        • V Vengatachalapathy Palanivel

                                          Switch case is better in performance than if-else-if ladder. While using switch case, compiler optimizes the code for the fastest execution. you can check this Url also: http://www.blackwasp.co.uk/SpeedTestIfElseSwitch.aspx[^]

                                          Regards, Vengat P

                                          M Offline
                                          M Offline
                                          Mark Salsbery
                                          wrote on last edited by
                                          #28

                                          That was my point[^]

                                          Mark Salsbery Microsoft MVP - Visual C++ :java:

                                          1 Reply Last reply
                                          0
                                          Reply
                                          • Reply as topic
                                          Log in to reply
                                          • Oldest to Newest
                                          • Newest to Oldest
                                          • Most Votes


                                          • Login

                                          • Don't have an account? Register

                                          • Login or register to search.
                                          • First post
                                            Last post
                                          0
                                          • Categories
                                          • Recent
                                          • Tags
                                          • Popular
                                          • World
                                          • Users
                                          • Groups