Add/Remove data object from dataset

Question

Hello,
I was working out a way to add and remove data from a dataset
on the fly and could not figure out any built in methods to do
this.
As I have decided to include spry in my project I like to try
and utilise as much as of its code as possible since it is
complicated to explain I have created a simple example – a
colour picker! (thought it might be more interesting…) of
what I am trying to achieve at
http://www.freshfresh.co.uk/spry/
- if you have a go on this and maybe look at the source code (all
the JS is embedded in HTML there are no modifications to other the
other core files). I have only used spry effects etc, including the
‘accordion’ – which I have become quite attached
to, it is really good for condensing pages down.
You will see I am using the setDataFromDoc method to create
the dataset – I tried the .data = myArrayOfData; .dataHash =
hashTable; method i.e. creating from an object rather than array
but it did’nt seem to play ball with the
addDataChangedObserver method – it did’nt update itself
each time it was modified (I am sorry I cannot remember exactly
what I did – but I tried all kinds of ways…). I stuck
with the string method because it worked - each time my new dataset
changed it updated itself on the screen (you will have to have a
look to understand.....sorry!!), but I would be interested to know
how such a thing could be implemented using the object route if you
believe this would be more efficient.
My second question is more simple – is there a shorter
way to add and remove a data object from a dataset? – as you
can see from the source code I have effectively created an
‘interface’ to do these tasks…. I could’nt
work out whether these methods are already built in. Maybe they
are?
Third question is… to extract a data object from a
dataset I use the .dataHash[the_row_id] method – is this the
right thing to do or could it lead to complications… that is
using methods that might supposedly be private?
Fourth question (observation really) – whilst creating
that colour picker example I went over board with my datasets and
loaded in several palettes some of which had over 1000 elements or
data objects. Which inevitably was very slow (on my computer
anyway) – but it got me thinking about trimming the contents
of my spry regions for better performance. I am I right in thinking
that the less HTML etc that there is in a spry region the quicker
SPRY will process it. For example say you had spry repeat with an
image tag in with some onclick, onmousover, onmouseout, style
attributes etc and compare this to a spry repeat with a simple
image with minimal attributes set. I suppose what I am trying to
say is – does spry ‘store’ all the contents of
each spry region somewhere? Or does it just process it and leave it
to the browser dom?
Fifth observation.... I find it really difficult to explain
computer technicalities in writing. It must be tough reading these
posts.... I know I find it difficult sometimes when dealing with
written end user feedback!
Andrew

Answer 1

Just clarifying my questions a bit further....
I found some old code regarding question 1 by using the
object method I mean something like this...
var mySwatches = [{'@hex':'ff0000'},{'@hex':'00ff00'}];
var hashTable = [];
function createDs(){
for (var i = 0; i < mySwatches.length; i++)
mySwatches
.ds_RowID = i;
hashTable = mySwatches
dsMySwatches.data = mySwatches;
dsMySwatches.dataHash = hashTable;
dsMySwatches.loadData();
i.e. not writing out a whole XML string string as the online
example does. When using this way I did'nt seem to be able to get
the HTML to refresh. I tried using [
Spry.Data.updateRegion('mydata'); ] after recreating the dataset I
also tried adding an [ .addDataChangedObserver ] (like in the
string example) amongst numerous other ways but it just would not
work like the string way. - Maybe I did something wrong somewhere.
In question 3 I refer to the [ .hash ] method .... its not a
'method' its a 'property' - my question should read - is it ok to
access private properties (from a browser campatiblity/security
point of view) that do not have specific methods to gain access to
them. I suppose it does'nt really matter with JS...
In question 4 I mention minimising the amount of code in a
spry region to speed it up. A clearer example of this might be for
example - a gallery with lots of images. As we know there will be a
slight delay as SPRY writes all the html so to speed up that intial
write I strip out all the image attributes such as onlclick do
this, onmouse over do that... and add these after the images have
loaded using a seperate function similar to my
fillSwatches(ds,prefix) function in my online colorpicker example.
I suppose it like a 2 tier processing of all the data. SPRY does
the intial display writing to get everything in place and then
another pass is made over to add any further functionality
adjustments etc. I am still not sure if that makes any sense!
***edit
Also on the subject of speed and the application as a whole
i.e. including my PHP - In one example I was creating I ended up
with an XML structure where each node has over 14 attributes i.e.
<somenode att1=”x” ……..
att14=”z”/> - as the file grew it obviously took
longer to process particularly on the server side, i.e. added all
those attributes just slowed it all down. So I did
this…… <somenode att1=”x:y:z” /> i.e
condensed selected attributes into a string that I could explode
later on.
Obviously this limits SPRYS ability to access the attributes
using the {attr} syntax. I had to create a function to explode the
array and do the ‘necessary’ on a second pass over the
data – this works ok for me. But its interesting that in this
particular case the server could not refresh the XML in an
acceptable time without doing this – just thought that might
be interesting to you. I suppose technically what I am doing is
abusing the concept of an XML structure and simply using it as a
‘carrier’ to feed my application …. Which I
suppose is where JSON comes in…. which is a bit more compact
and maybe faster to manipulate on both the server and client side
– I don’t really know, I have never used it –
just throwing ideas around!!
Andrew