ignite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dmitriy Setrakyan <dsetrak...@apache.org>
Subject Re: Cache.invoke(...) vs GridCache.transform() methods
Date Mon, 19 Jan 2015 08:46:44 GMT
Good to know. Looks like no work needs to be done here. Have we also
migrated the tests for "transform" methods to use "invoke" methods?

D.

On Mon, Jan 19, 2015 at 12:35 AM, Semyon Boikov <sboikov@gridgain.com>
wrote:

> Currrent implementation of Cache.invoke is slightly changed implementation
> of transform(), so Cache.invoke supports transactions.
>
> On Sun, Jan 18, 2015 at 3:49 AM, Dmitriy Setrakyan <dsetrakyan@apache.org>
> wrote:
>
> > I had a discussion with Alexey G. today and it turns out that we may have
> > missed an important use case that was available with
> > GridCache.transform(...) method (which was removed) and is no longer
> > supported with Cache.invoke(...).
> >
> > The difference is that Cache.invoke(EntryProcessor, ...) does everything
> > GridCache.transform(...) could do, except transactions.
> > GridCache.transform(...) methods, however, can be invoked within
> > transactions.
> >
> > Is it possible to support transactions with Cache.invoke(EntryProcessor,
> > ...) methods?
> >
> > Or should we bring the transform(..) methods back? If yes, then they
> should
> > look like this:
> > ----
> > <T> T transform(K, IgniteClosure<T, MutableEntry<K, V>)
> > <T> T transformAll(Collection<K>, IgniteClosure<T, MutableEntry<K,
V>>)
> > ----
> >
> > Thoughts? (especially Alexey G, and Semyon, since you were working on
> this
> > feature)
> >
> > D.
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message