make array empty

Discussion in 'Javascript' started by Sunny, Oct 8, 2008.

  1. Sunny

    Sunny Guest

    Hi, Can someone tell me, How to redefine array or make array empty or
    null.

    Here what I am trying to do.
    var temp = new Array();
    for(i=0; i <=outstring.length-1; i++) {
    temp = outstring.split(',');
    }

    Once the for loop will finish doing it, I want temp array to be null
    or blank.
    How Can I do that?
    I tried doing temp.length =0 but that didn't work.
     
    Sunny, Oct 8, 2008
    #1
    1. Advertising

  2. Sunny

    Laser Lips Guest

    On Oct 8, 3:10 pm, Sunny <> wrote:
    > Hi, Can someone tell me, How to redefine array or make array empty or
    > null.
    >
    > Here what I am trying to do.
    > var temp = new Array();
    > for(i=0; i <=outstring.length-1; i++) {
    > temp = outstring.split(',');
    >
    > }
    >
    > Once the for loop will finish doing it, I want temp array to be null
    > or blank.
    > How Can I do that?
    > I tried doing temp.length =0 but that didn't work.


    temp=[];
     
    Laser Lips, Oct 8, 2008
    #2
    1. Advertising

  3. Sunny

    RobG Guest

    On Oct 9, 12:42 am, Conrad Lender <> wrote:
    [...]
    >
    >   var temp = [];   // unless see below [*]

    [...]
    >
    > [*] That is, unless you're going push() a lot of elements on the array,
    > and you know the final number in advance; then it's more efficient to
    > use "new Array(num_elements)". I'm not exactly sure why that is, because
    > the implementations don't reserve memory in advance; maybe it's that the
    > .length property doesn't change after each push().


    Have you compared that to using a while loop? e.g.

    var t = [];
    var i = array.length;

    while (i--) {
    t.push(array);
    }

    That should only set length once too, but I imagine the internal
    [[put]] method must still check whether length needs to be
    incremented, even if it doesn't have to actually do it.


    --
    Rob
     
    RobG, Oct 9, 2008
    #3
  4. Sunny

    sasuke Guest

    On Oct 8, 7:42 pm, Conrad Lender <> wrote:
    > On 2008-10-08 16:10, Sunny wrote:
    >
    > > Hi, Can someone tell me, How to redefine array or make array empty or
    > > null.

    >
    > > Here what I am trying to do.
    > > var temp = new Array();
    > > for(i=0; i <=outstring.length-1; i++) {
    > > temp = outstring.split(',');
    > > }

    >
    > I assume you're doing something else in the loop too, or this would be
    > pretty pointless.
    >
    > First of all, you don't need
    >
    >   var temp = new Array();
    >
    > because you'll immediately assign something else to temp in the loop.
    >
    >   var temp;
    >
    > will do. By the way, if you do want to create an empty array, use an
    > array literal:
    >
    >   var temp = [];   // unless see below [*]
    >
    > > Once the for loop will finish doing it, I want temp array to be null
    > > or blank.

    >
    > Why? Just let it go out of scope, and it will be garbage collected
    > (unless there are closures present).
    >
    > > How Can I do that?
    > > I tried doing temp.length =0 but that didn't work.

    >
    > Are you really sure about that?
    > How didn't it work?
    >
    > Alternatively, you could also use the delete operator, or you could
    > assign something else to temp, if you want, like null or [].
    >
    >   - Conrad
    >
    > [*] That is, unless you're going push() a lot of elements on the array,
    > and you know the final number in advance; then it's more efficient to
    > use "new Array(num_elements)". I'm not exactly sure why that is, because
    > the implementations don't reserve memory in advance; maybe it's that the
    > .length property doesn't change after each push().


    IMO, memory is reserved for the data structure which is responsible
    for maintaining the state of the Array object; it can be an Object[]
    of Java or an array of void pointers in C. The advantage AFAIK here is
    that setting the 'length' property avoids a lot of copying of
    references to and fro when inserting an element since each time the
    backing data store has to be expanded to take in the new element i.e.
    ensureCapacity does nothing as long as the number of elements
    sequentially inserted is less than the specified length.

    Even I think that setting the length property to 0 should have done
    the job though setting it to null or just letting it go out of scope
    of more convenient and processing friendly [since setting the length
    or deleting actually loops over the array elements].
     
    sasuke, Oct 9, 2008
    #4
  5. On Wed, 8 Oct 2008 at 22:53:26, in comp.lang.javascript, sasuke wrote:

    <snip>
    >IMO, memory is reserved for the data structure which is responsible
    >for maintaining the state of the Array object; it can be an Object[]
    >of Java or an array of void pointers in C. The advantage AFAIK here is
    >that setting the 'length' property avoids a lot of copying of
    >references to and fro when inserting an element since each time the
    >backing data store has to be expanded to take in the new element i.e.
    >ensureCapacity does nothing as long as the number of elements
    >sequentially inserted is less than the specified length.


    It can't be as simple as that. Try putting this into the address bar and
    executing it :

    javascript: var a1 = new Array(); a1[2000000000] = "Wow"; alert(a1.length);

    It doesn't blow up. Neither does this :

    javascript: var a1 = new Array(2000000000); alert(a1.length );


    >Even I think that setting the length property to 0 should have done
    >the job though setting it to null or just letting it go out of scope
    >of more convenient and processing friendly [since setting the length
    >or deleting actually loops over the array elements].


    According to ECMA 262, setting the length to zero should indeed do the job :

    "whenever the length property is changed, every property whose
    name is an array index whose value is not smaller than the new length is
    automatically deleted."

    As you say, replacing the variable's value by a new array or null will also do
    the job. Whether it is quicker depends on how the garbage collector works.

    John
    --
    John Harris
     
    John G Harris, Oct 9, 2008
    #5
  6. Sunny

    sasuke Guest

    On Oct 10, 12:41 am, John G Harris <> wrote:
    > On Wed, 8 Oct 2008 at 22:53:26, in comp.lang.javascript, sasuke wrote:
    >
    >   <snip>
    >
    > >IMO, memory is reserved for the data structure which is responsible
    > >for maintaining the state of the Array object; it can be an Object[]
    > >of Java or an array of void pointers in C. The advantage AFAIK here is
    > >that setting the 'length' property avoids a lot of copying of
    > >references to and fro when inserting an element since each time the
    > >backing data store has to be expanded to take in the new element i.e.
    > >ensureCapacity does nothing as long as the number of elements
    > >sequentially inserted is less than the specified length.

    >
    > It can't be as simple as that. Try putting this into the address bar and
    > executing it :
    >
    >   javascript: var a1 = new Array(); a1[2000000000] = "Wow"; alert(a1.length);
    >
    > It doesn't blow up. Neither does this :
    >
    >   javascript: var a1 = new Array(2000000000);  alert(a1.length );


    My apologies, I should have been more explicit. Though there is no
    mention of such in the specification, almost all implementations have
    a `sparse' or `dense' flag which is used to mark the Array object.
    This flag demands a suitable processing on part of the implementation
    when a high value of index is used; which explains your first e.g.

    As for the second e.g., try something like javascript: var a1 = new
    Array(2000000000); for(var i = 0, maxI = a1.length; i < maxI; ++i)
    { a1 = Number(i); }

    The script stops responding, doesn't it? This gives us an indication
    that the implementation is smart enough to allocate memory to the
    Array object when it actually is required [maybe the length passed to
    the Array constructor isn't used till a reference to a particular
    location isn't made].

    > >Even I think that setting the length property to 0 should have done
    > >the job though setting it to null or just letting it go out of scope
    > >of more convenient and processing friendly [since setting the length
    > >or deleting actually loops over the array elements].

    >
    > According to ECMA 262, setting the length to zero should indeed do the job :
    >
    > "whenever the length property is changed, every property whose
    > name is an array index whose value is not smaller than the new length is
    > automatically deleted."
    >
    > As you say, replacing the variable's value by a new array or null will also do
    > the job. Whether it is quicker depends on how the garbage collector works..


    Yes, what needs to be done greatly depends on the implementation
    details which makes the question "which is the fastest way to reclaim
    memory taken by the Array xxx" way too generic or difficult to answer.

    /sasuke
     
    sasuke, Oct 10, 2008
    #6
  7. Sunny

    dhtml Guest

    John G Harris wrote:
    > On Wed, 8 Oct 2008 at 22:53:26, in comp.lang.javascript, sasuke wrote:



    ECMAScript does not have real arrays. Arrays are objects.

    >
    > It can't be as simple as that. Try putting this into the address bar and
    > executing it :
    >
    > javascript: var a1 = new Array(); a1[2000000000] = "Wow"; alert(a1.length);
    >


    That creates a new Array with two properties: 2000000000 and length.

    2000000000 = "Wow"
    length = 2000000001

    > It doesn't blow up. Neither does this :
    >
    > javascript: var a1 = new Array(2000000000); alert(a1.length );
    >
    >


    That creates a new Array with one property: length.

    Firefox <= 3.0.3 exhibits a bug where an array is 'prefilled' with
    properties having undefined values.

    javascript:alert('2' in [,,,,,,,,,,])

    true in firefox 3.0.3

    (should be false).


    Garrett

    >
    > John
     
    dhtml, Oct 10, 2008
    #7
  8. sasuke wrote:
    > [...] Though there is no mention of such in the specification, almost all
    > implementations have a `sparse' or `dense' flag which is used to mark the
    > Array object. This flag demands a suitable processing on part of the
    > implementation when a high value of index is used; which explains your
    > first e.g. [...]


    How do you got *that* idea?

    > As for the second e.g., try something like javascript: var a1 = new
    > Array(2000000000); for(var i = 0, maxI = a1.length; i < maxI; ++i) {
    > a1 = Number(i); }
    >
    > The script stops responding, doesn't it?


    Responding to what?

    > This gives us an indication that the implementation is smart enough to
    > allocate memory to the Array object when it actually is required [maybe
    > the length passed to the Array constructor isn't used till a reference to
    > a particular location isn't made].


    Your logic is flawed. The *user agent* stops responding (and may show the
    user a dialog that allows them to stop execution of the script) because the
    script is running quite a long time and all known ECMAScript implementations
    are single-threaded. It is completely irrelevant that an Array object is
    involved here. Simple proof:

    while (true);


    PointedEars
    --
    Use any version of Microsoft Frontpage to create your site.
    (This won't prevent people from viewing your source, but no one
    will want to steal it.)
    -- from <http://www.vortex-webdesign.com/help/hidesource.htm>
     
    Thomas 'PointedEars' Lahn, Oct 11, 2008
    #8
  9. Sunny

    sasuke Guest

    On Oct 11, 4:51 am, Thomas 'PointedEars' Lahn <>
    wrote:
    > sasuke wrote:
    > > [...] Though there is no mention of such in the specification, almost all
    > > implementations have a `sparse' or `dense' flag which is used to mark the
    > > Array object. This flag demands a suitable processing on part of the
    > > implementation when a high value of index is used; which explains your
    > > first e.g. [...]

    >
    > How do you got *that* idea?


    When sifting through the source code of Rhino; the Java implementation
    of ECMAScript. The NativeArray class maintains a flag called
    'denseOnly'.

    > > As for the second e.g., try something like javascript: var a1 = new
    > > Array(2000000000);  for(var i = 0, maxI = a1.length; i < maxI; ++i) {
    > > a1 = Number(i); }

    >
    > > The script stops responding, doesn't it?

    >
    > Responding to what?
    >
    > > This gives us an indication that the implementation is smart enough to
    > > allocate memory to the Array object when it actually is required [maybe
    > > the length passed to the Array constructor isn't used till a reference to
    > >  a particular location isn't made].

    >
    > Your logic is flawed.  The *user agent* stops responding (and may show the
    > user a dialog that allows them to stop execution of the script) because the
    > script is running quite a long time and all known ECMAScript implementations
    > are single-threaded.  It is completely irrelevant that an Array object is
    > involved here.  Simple proof:
    >
    >   while (true);


    I concur; I used a wrong example in trying to put forth my
    observation. Another small experiment I tried was tracking the memory
    usage when running the examples:

    Case 1:
    javascript: var a = new Array(200000000); for(var i = 0; i < 5000; +
    +i) { a[i * 10] = i; }

    Case 2:
    javascript: var a = new Array(200000000); for(var i = 0; i < 5000; +
    +i) { var j = i + 1; }

    Case 3:
    javascript: var a = new Array(200000000); a[1000] = 1000;

    The observations were that when running e.g. 1, there was a sharp
    memory spike of the order of several KB's which wasn't seen when
    running e.g. 2 and 3.

    I agree that these small tests don't prove anything; just an
    observation, that's it.
     
    sasuke, Oct 11, 2008
    #9
  10. sasuke wrote:
    > Thomas 'PointedEars' Lahn wrote:
    >> sasuke wrote:
    >>> [...] Though there is no mention of such in the specification, almost all
    >>> implementations have a `sparse' or `dense' flag which is used to mark the
    >>> Array object. This flag demands a suitable processing on part of the
    >>> implementation when a high value of index is used; which explains your
    >>> first e.g. [...]

    >> How do you got *that* idea?

    >
    > When sifting through the source code of Rhino; the Java implementation
    > of ECMAScript. The NativeArray class maintains a flag called
    > 'denseOnly'.


    While that is interesting, this is but one implementation, the wrong one for
    the discussed example (we are talking SpiderMonkey and friends in this
    thread), and it does not explain the example (whereas it is unclear what
    exactly needed explaining there).


    PointedEars
    --
    realism: HTML 4.01 Strict
    evangelism: XHTML 1.0 Strict
    madness: XHTML 1.1 as application/xhtml+xml
    -- Bjoern Hoehrmann
     
    Thomas 'PointedEars' Lahn, Oct 11, 2008
    #10
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. John

    empty/non-empty element

    John, Jul 15, 2003, in forum: XML
    Replies:
    1
    Views:
    1,066
    Klaus Johannes Rusch
    Jul 16, 2003
  2. Lukas
    Replies:
    3
    Views:
    868
    spiff
    Nov 10, 2005
  3. Marcia Hon

    Check if a directory is empty and empty it

    Marcia Hon, Feb 8, 2004, in forum: C Programming
    Replies:
    8
    Views:
    614
    Dave Thompson
    Feb 14, 2004
  4. Brian Roberts

    empty lists vs empty generators

    Brian Roberts, May 3, 2005, in forum: Python
    Replies:
    12
    Views:
    683
    Jeremy Bowers
    May 4, 2005
  5. ButlerDJIAM
    Replies:
    0
    Views:
    580
    ButlerDJIAM
    Nov 9, 2006
Loading...

Share This Page