Alexey, would should our users even know about the existence of _key and
_val fields? These fields, if needed, should be completely internal. Do you
disagree?

On Wed, Mar 15, 2017 at 5:32 PM, Alexey Kuznetsov <akuznet...@apache.org>
wrote:

> Dmitriy,
>
> Are you sure???
>
> AFAIK _key and _val fields will be declared as "invisible" i.e. will be not
> included in "select * from T" queries,
> but NOT removed and still could be used.
>
> Or please give a link to JIRA issue.
>
>
> On Thu, Mar 16, 2017 at 7:27 AM, Dmitriy Setrakyan <dsetrak...@apache.org>
> wrote:
>
> > I completely agree. Moreover, AFAIK the _key and _val fields will be
> > removed in 2.0, so we must add ID or any other field used for a primary
> key
> > to the class.
> >
> >
> > On Wed, Mar 15, 2017 at 5:19 PM, Denis Magda <dma...@apache.org> wrote:
> >
> >> Alex K., Igniters,
> >>
> >> It’s is stated that an existing JDBC based application that works with a
> >> relation database should be easily migrated to Apache Ignite if we used
> Web
> >> Console and JDBC driver. So, I decided to prove this.
> >>
> >> A simple application was created that issues a bunch of SELECT, UPDATE,
> >> etc. queries.
> >>
> >> For instance, the simplest query I had looks like this:
> >>
> >> SELECT * FROM city WHERE id = ?
> >>
> >> When Web Console is used for the schema importing purposes it doesn’t
> add
> >> ‘id’ field to a City POJO and, consequently, doesn’t define index for
> this
> >> field. *From technical perspective*, I do understand why we do this - to
> >> save some memory because cache entries key will be used as ‘id’ instead.
> >>
> >> But this breaks the usability in some sense because my simplest query
> >> above no longer works.
> >>
> >> What if we add a special checkbox to Web Console that will add fields
> >> like ‘id’ to a POJO? If the checkbox is disabled (default) we should
> show a
> >> message at the final importing stage saying that fields like ‘id’ has
> to be
> >> replaced with ‘_key’.
> >>
> >> Thoughts?
> >
> >
> >
>
>
> --
> Alexey Kuznetsov
>

Reply via email to