[jbosscache-dev] Node.getChildren()

[Ben Wang]

For CachedMaplImpl, I have been relying on keySet() and entrySet() from the returned Map to get Fqn, Node Collection. Yes, I can create it myself from the returned Collection by looping. But it will do the extra work from your fresh Collection instance, will it not?


-----Original Message-----
From: Manik Surtani [mailto:manik at jboss.org] 
Sent: Tuesday, September 05, 2006 12:06 AM
To: Ben Wang
Cc: jbosscache-dev at lists.jboss.org
Subject: Re: [jbosscache-dev] Node.getChildren()

So if you don't know the Fqn in advance, how does having a Map<Fqn,  
Node> help over having a Collection<Node>?  The only benefit I can
see here is doing Map<Fqn, Node>.containsKey()... all else should be just as easy/efficient with Collection<Node>, IMO.  Am I missing something?


--
Manik Surtani

Lead, JBoss Cache
JBoss, a division of Red Hat

Email: manik at jboss.org
Telephone: +44 7786 702 706
MSN: manik at surtani.org
Yahoo/AIM/Skype: maniksurtani


On 4 Sep 2006, at 17:01, Ben Wang wrote:

> In my case, I am querying for all children Node and find out their 
> respective (fqn, Node). Problem is that I don't know the fqn in 
> advance. I am using this currently for Collection implementation for 
> keySet and entrySet, for example.
>
> Cheers,
>
> -Ben
>
> -----Original Message-----
> From: Manik Surtani [mailto:manik at jboss.org]
> Sent: Monday, September 04, 2006 7:14 PM
> To: Ben Wang
> Cc: jbosscache-dev at lists.jboss.org
> Subject: Re: [jbosscache-dev] Node.getChildren()
>
> Hi Ben.
>
> Again, this (rather poor) implementation was just to get the API 
> contracts in place.  It does need refactoring in the way it is 
> implemented, but I see what you mean about getChildren() returning a 
> Map<Fqn, Node> being more useful than returning a Collection<Node>.
> The latter (and current impl) is cleaner from an OO perspective, but 
> if the former is more practical then so be it.
>
> Just so I understand the use case better, why do you need a Map<Fqn,
> Node>?  If you are looking for a specific child, wouldn't getChild
> (Fqn) suffice?
>
> Cheers,
> --
> Manik Surtani
>
> Lead, JBoss Cache
> JBoss, a division of Red Hat
>
> Email: manik at jboss.org
> Telephone: +44 7786 702 706
> MSN: manik at surtani.org
> Yahoo/AIM/Skype: maniksurtani
>
>
> On 4 Sep 2006, at 04:32, Ben Wang wrote:
>
>> Manik,
>>
>> I have noticed that this is the current implementation for
>> Node.getChildren():
>>
>>
>>    public Collection<Node> getChildren()
>>    {
>>       Map m = currentNode.getChildren();
>>       Set<Node> children = new HashSet<Node>(m.size());
>>       Iterator i = m.values().iterator();
>>       while (i.hasNext())
>>       {
>>          children.add(new TreeCacheProxyImpl(treeCache, (NodeImpl) 
>> i.next()));
>>       }
>>
>>       return children;
>>    }
>>
>>
>> Actually, what I need now is a map of (keyFqn, node), i.e., return m. 
>> But even if we stick to this api, creating children will be quite 
>> expensive won't it be if getChildren is called many times (happens 
>> for my case dealing with Collection classes).
>>
>> Thanks,
>>
>> -Ben
>>
>> _______________________________________________
>> jbosscache-dev mailing list
>> jbosscache-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbosscache-dev
>