Variable verses Array Difference?

[Pured]

Hello,

I just wanted to know why variables don't need to be explicitly defined as "global/local" and arrays do. I've read that arrays need it, but not why.

 

Thank you in advance.

[Luigi]

I thinking becouse is diferent data types, one is 'array', other is the rest (string, int, etc).

I can't anymore write a code whitout local or global.

And I always use:

Opt("MustDeclareVars", 1)

It force declare all variables with local or global.

Define a variable with DIM, it's old. I do not use.

If is a small code, no problem.

If the code have more than 100 lines... have so much variables, I prefer declare.

 

Edited September 10, 2017 by Luigi

[Pured]

@Luigi, I've considered using Opt(), but I know where I'm defining my variables, so it's not too much of an issue. I guess I'll just start using global/local for all variables regardless of how obvious it's scope is.

[HurleyShanabarger]

21 minutes ago, Pured said:

@Luigi, I've considered using Opt(), but I know where I'm defining my variables, so it's not too much of an issue. I guess I'll just start using global/local for all variables regardless of how obvious it's scope is.

It will matter once you will have a script where you do care about the speed. So better do it from the beginning and establish a good way of coding,

[Pured]

@Hurley, Thanks. I never thought about its impact on speed.

[iamtheky]

they dont require anything special, see:

the return of any function at all that returns an array and Dim.  And scoping isnt going to help with anything at all except maintenance of the code.  Also, there is magic that happens behind the scenes:

 

Edited September 10, 2017 by iamtheky

[Pured]

@iam, Okay, so speed isn't actually affected since the compiler defines them before runtime, correct?

[iamtheky]

there are like 2 people max that can give you an answer that isnt borderline retarded (or at least a blind guess).  I am not one of those people.  However, I would never accept a speed argument without a Proof of Concept.

[HurleyShanabarger]

11 minutes ago, Pured said:

@iam, Okay, so speed isn't actually affected since the compiler defines them before runtime, correct?

Judge by yourself - I know the differnce is mininmal but think in bigger dimensions:

expandcollapse popup

Global $giTest
Global $giTest_ByRef
$iLoopMax = 5000000

$iTimer = TimerInit()
For $iLoop = 0 To $iLoopMax
    _Main_V1()
Next
ConsoleWrite(TimerDiff($iTimer) & @CRLF)

$iTimer = TimerInit()
For $iLoop = 0 To $iLoopMax
    _Main_V2()
Next
ConsoleWrite(TimerDiff($iTimer) & @CRLF)

$iTimer = TimerInit()
For $iLoop = 0 To $iLoopMax
    _Main_V3()
Next
ConsoleWrite(TimerDiff($iTimer) & @CRLF)


$iTimer = TimerInit()
For $iLoop = 0 To $iLoopMax
    _Main_V4($giTest_ByRef)
Next
ConsoleWrite(TimerDiff($iTimer) & @CRLF)

Func _Main_V1()
    $giTest = Random()
    Return $giTest
EndFunc

Func _Main_V2()
    $iTest = Random()
    Return $iTest
EndFunc

Func _Main_V3()
    Local $iTest = Random()
    Return $iTest
EndFunc

Func _Main_V4(ByRef $iInt)
    $iInt = Random()
EndFunc

//Edit: So more or less I was really wrong.

//Edit2: At least when it comes to usage of memory it might be better if you use a local scope.

Edited September 10, 2017 by HurleyShanabarger
Stupid script :D

[iamtheky]

I ran the test 3 times, with the functions in 3 different orders.  Tell me if you can identify the trend?

;~ 0.00950340526017283 - v1
;~ 0.008743132839359 - v2
;~ 0.00380136210406913 - v3

;~ 0.0380136210406913  - v2
;~ 0.022047900203601 - v3
;~ 0.0167259932579042 - v1

;~ 0.0106438138913936 - v3
;~ 0.00532190694569678 - v1
;~ 0.00380136210406913 - v2

 

[HurleyShanabarger]

5 minutes ago, iamtheky said:

I ran the test 3 times, with the functions in 3 different orders.  Tell me if you can identify the trend?

;~ 0.00950340526017283 - v1
;~ 0.008743132839359 - v2
;~ 0.00380136210406913 - v3

;~ 0.0380136210406913  - v2
;~ 0.022047900203601 - v3
;~ 0.0167259932579042 - v1

;~ 0.0106438138913936 - v3
;~ 0.00532190694569678 - v1
;~ 0.00380136210406913 - v2

 

Edit my script - there was a mistake inside; see my edit.

[iamtheky]

still no matter which way i run them, no matter what else is going on with my system, they take progressively longer to execute.  Either the test is not accurate or back to my hypothesis, scoping does not matter in relationship to speed.

;~ 6990.2510267479 - v1
;~ 11745.9158165554 - v2
;~ 11969.2507819402 - v3
;~ 12215.0099823769 - v4


;~ 16821.4408987028 - v4
;~ 21365.0269364519 - v3
;~ 18593.8989658775 - v2
;~ 18698.0642703894 - v1

;~ 3788.5214830178 - v3
;~ 5206.71693081065 - v4
;~ 5446.71326629758 - v1
;~ 5676.78234464974 - v2

 

Edited September 10, 2017 by iamtheky