I would be very wary of assuming that the for loop goes Last In First
Out. To quote the Mozilla docs (http://developer.mozilla.org/en/docs/
Core_JavaScript_1.5_Reference:Statements:for...in):
A for...in loop iterates over the properties of an object in an
arbitrary order.
It may work correctly in the browsers you tested, but could break at
any time.
Danny Wachsstock
On Oct 11, 3:30 pm, Flesler <[EMAIL PROTECTED]> wrote:
> The for iteration goes from the older ones to newer ones, so if you
> want to keep newer ones (FIFO):
> function pruneCache(amt) {
> for (var i in cache) {
> if (cache.length <= amt) return;//pruned
> delete cache[i];
> cache.length--;
> }}
>
> that asumes you are keeping the length updated till here. If you want
> to remove newer first(LIFO):
> function pruneCache(amt) {
> delete cache.length;//remove length to avoid iterating it
> var copy = {length:0}; //new hash
> for (var i in cache) {
> if (copy.length >= amt) break;//done
> copy[i] = cache[i];
> copy.length++;
> }
> cache = copy;//replace the old cache
>
> }
>
> The other option is to have 2 parallel arrays of keys and values...
> Hope that helps.
>
> Ariel Flesler
>
> On Oct 11, 4:54 pm, Pops <[EMAIL PROTECTED]> wrote:
>
> > Thanks Mike and Wizzud.
>
> > Question:
>
> > What I wanted to use this for was my cache and to "truncate" old
> > data. I see that using a real array will allow me to use the
> > inherited .length property to set a new size and it will do
> > truncation. But since I am using an associated array, the length
> > property is no longer usable.
>
> > So if I use the for loop with delete, will I be pulling the "rug from
> > under its feet?"
>
> > Typically, in code designs like this, you would do a reverse traversal
> > to delete the last entries first so you can keep with the internal
> > loop counters and/or references.
>
> > In other words, is this "safe?"
>
> > function pruneCache(amt) {
> > var n = 0;
> > for (var i in cache) {
> > if (amt > 0) {
> > amt--;
> > delete cache[i];
> > return;
> > }
> > n++;
> > }
> > cache.length = n;
> > }
>
> > If not, then in lieu of a reverse loop syntax, I would probably need
> > to copy the cache first? I have not checked but if returning false
> > stops the traveral that would be more efficient.
>
> > I could recode all this into a pure indexed array and then use
> > the .length, but then I would need a fast lookup method or 2nd matrix
> > to map the associated name with the array index.
>
> > Thanks
>
> > --
> > HLS
>
> > On Oct 11, 4:26 am, "Michael Geary" <[EMAIL PROTECTED]> wrote:
>
> > > > From: Pops
>
> > > > Is this the only way to iterate through an object?
>
> > > > for ( x in myObject) {
> > > > }
>
> > > Ultimately, yes. You can use something like $.each() on an object, but it
> > > just runs that same for loop internally - see the source code for
> > > $.each().
>
> > > > and how can you remove an field in an object?
>
> > > > For example:
>
> > > > var obj = {
> > > > x1: 1123,
> > > > x2: 212,
> > > > x3: 123131
> > > > };
>
> > > > I want to remove obj.x2 the obj object to end up with:
>
> > > > { x1: 1123, x3: 123131}
>
> > > delete obj.x2;
>
> > > or
>
> > > delete obj['x2'];
>
> > > Note that it is not an error if the property doesn't exist, so you don't
> > > have to "protect" the delete with code like this:
>
> > > if( 'x2' in obj ) delete obj.x2; // unnecessary precaution
>
> > > -Mike- Hide quoted text -
>
> > - Show quoted text -
