You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@phoenix.apache.org by Dan Di Spaltro <da...@gmail.com> on 2014/06/12 06:47:49 UTC

setReadOnly

I have a sql accessing library that I was hoping to use for phoenix, but
they use readOnly connections to get the metadata.  As it stands Phoenix
throws an exception when you set a conn as readOnly.  Do you think that
would be reasonable to change this to a warning or remove the exception
entirely.  I get the reasons not to, but I also don't want to change a
bunch of code or keep a separate branch on something that would actually
work, just not be 100% truthful.

Thoughts?

-- 
Dan Di Spaltro

Re: setReadOnly

Posted by Dan Di Spaltro <da...@gmail.com>.
Sorry I missed this, thanks for the response.  Yeah I worked around it for
now, sometimes these sql access layers set things that make sense but are
out of your control.  That seems like a good nuclear option, but maybe if
no one else cares Ill just stick with what I have.  BTW, I was using this
library http://scalikejdbc.org/  It sets readOnly for asking about the
metadata of each table...

-Dan


On Thu, Jun 12, 2014 at 8:19 AM, Gabriel Reid <ga...@gmail.com>
wrote:

> Hi Dan,
>
> I've encountered a similar scenario with
> Connection.setTransactionIsolation, which also throws a
> SQLFeatureNotSupportedException. On the one hand, it's definitely
> annoying to have little things like this getting in the way, but on
> the other hand I'm not crazy about the idea of just logging a warning
> for these features. An application may be doing things that are
> actually dependent on these settings working, so I prefer the idea of
> failing fast.
>
> One idea might be to have a connection setting called
> "ignoreUnsupportedFeatures" or something like that, and if it was set
> to true then a warning could be logged instead of throwing
> SQLFeatureNotSupportedException in case something like setReadOnly was
> called. I think that would give a good balance of allowing getting
> around this issue without sacrificing too much in correctness.
>
> What do you think?
>
> - Gabriel
>
>
> On Thu, Jun 12, 2014 at 6:47 AM, Dan Di Spaltro <da...@gmail.com>
> wrote:
> > I have a sql accessing library that I was hoping to use for phoenix, but
> > they use readOnly connections to get the metadata.  As it stands Phoenix
> > throws an exception when you set a conn as readOnly.  Do you think that
> > would be reasonable to change this to a warning or remove the exception
> > entirely.  I get the reasons not to, but I also don't want to change a
> bunch
> > of code or keep a separate branch on something that would actually work,
> > just not be 100% truthful.
> >
> > Thoughts?
> >
> > --
> > Dan Di Spaltro
>



-- 
Dan Di Spaltro

Re: setReadOnly

Posted by Gabriel Reid <ga...@gmail.com>.
Hi Dan,

I've encountered a similar scenario with
Connection.setTransactionIsolation, which also throws a
SQLFeatureNotSupportedException. On the one hand, it's definitely
annoying to have little things like this getting in the way, but on
the other hand I'm not crazy about the idea of just logging a warning
for these features. An application may be doing things that are
actually dependent on these settings working, so I prefer the idea of
failing fast.

One idea might be to have a connection setting called
"ignoreUnsupportedFeatures" or something like that, and if it was set
to true then a warning could be logged instead of throwing
SQLFeatureNotSupportedException in case something like setReadOnly was
called. I think that would give a good balance of allowing getting
around this issue without sacrificing too much in correctness.

What do you think?

- Gabriel


On Thu, Jun 12, 2014 at 6:47 AM, Dan Di Spaltro <da...@gmail.com> wrote:
> I have a sql accessing library that I was hoping to use for phoenix, but
> they use readOnly connections to get the metadata.  As it stands Phoenix
> throws an exception when you set a conn as readOnly.  Do you think that
> would be reasonable to change this to a warning or remove the exception
> entirely.  I get the reasons not to, but I also don't want to change a bunch
> of code or keep a separate branch on something that would actually work,
> just not be 100% truthful.
>
> Thoughts?
>
> --
> Dan Di Spaltro