In 1.0.2 there is:
// request data
requestRows: function(inRowIndex, inCount){
var row = inRowIndex || 0;
var params = {
start: row,
count: this.rowsPerPage,
query: this.query,
onBegin: dojo.hitch(this, "beginReturn"),
// onItem: dojo.hitch(console, "debug"),
onComplete: dojo.hitch(this, "processRows") // add to deferred?
}
// console.debug("requestRows:", row, this.rowsPerPage);
this.store.fetch(params);
},
requestRows: function(inRowIndex, inCount){
var row = inRowIndex || 0;
var params = {
start: row,
count: this.rowsPerPage,
query: this.query,
onBegin: dojo.hitch(this, "beginReturn"),
// onItem: dojo.hitch(console, "debug"),
onComplete: dojo.hitch(this, "processRows") // add to deferred?
}
// console.debug("requestRows:", row, this.rowsPerPage);
this.store.fetch(params);
},
I'm loading some large data that I need to transform. I don't want to do that in onBegin, because that
will use up too much memory; I want to transform each row. But because onItem is commented out, I'm going to have to extend ItemFileReadStore instead of just connecting to whatever would have been hitched in onItem. Arrghh!
Should I not even extend -- is there some nasty problem with onItem that I'll only discover after hours with FireBug? Or is onItem commented out by accident?
I just wish an attach point had been left in for it, for just this kind of filtering.
Thanks much,
/r:b:
You would have had to extend
You would have had to extend DojoData anyway - since onItem was connected to console.debug (which you probably don't what to connect to...)
There is no onItem in the 1.1 or svn code either - so I'm guessing that it was removed on purpose. I don't think that it's necessary.
I don't know exactly what you are looking at doing or transforming - but I'm guessing that you'll need to extend ItemFileReadStore (which is a normal thing to need to do anyway) - possibly overriding the _getItemsFromLoadedData function. That allows for you to do your transformation up-front. The other thing you can do is to override _fetchItems - which will allow you to do your transformation as your data is loaded.
As far as memory goes, either one of these would take up the same amount as connecting to the onItem event on the store. Since ItemFileReadStore uses dojo.data.util.simpleFetch, I don't think that connecting to onItem will save you any memory at all...all the data is loaded at once - using _fetchItems - and then it is passed to onItem one at a time...