On Fri, May 29, 2020 at 8:53 PM Kenneth Knowles wrote:
> -0, sorry I'm late. But I guess you are not sorry since everyone else
> agrees :-)
>
> I just don't know what this accomplishes. Most components correspond to a
> software component/codebase. The only components that do not correspond to
>
Did you create a jira about that already ?I will do the improvement on JdbcIO. Regards JBThanksRegards JBLe dim. 31 mai 2020 ? 11:25, Willem Pienaar a ?crit :Hi Reuven,To be clear, we already have this solved for BigQueryIO. I am hoping there is a similar solution for JdbcIO.Regards,WillemOn Sun,
It doesn't look to me like JdbcIO currently supports dynamic destinations.
I think it wouldn't be too hard to add this functionality. If you wanted to
help contribute this change to JdbcIO.java, I'm sure that we would be happy
to help guide you.
Reuven
On Sun, May 31, 2020 at 2:26 AM Willem
On 5/30/20 5:39 AM, Kenneth Knowles wrote:
Agree to delete them, though for different reasons. I think this code
comes from a desire to have methods that can be called on a DoFn
directly. And from reviewing the code history I think they are copied
in from another class. So that's why they are
Answers inline.
On 5/29/20 5:46 PM, Luke Cwik wrote:
To go back to your original question.
I would remove the static convenience methods in DoFnSignatures since
they construct a DoFnSignature and then throw it away. This
construction is pretty involved, nothing as large as an IO call but it
Minor self-correction - in the property c) it MIGHT be possible to
update output watermark to time greater than input watermark, _as long
as any future element cannot be assigned timestamp that is less than the
output watermark_. That seems to be the case only for
Hi Reuven,
the asynchronicity of watermark update is what I was missing - it is
what relates watermarkhold with element output timestamp. On the other
hand, we have some invariants that have to hold, namely:
a) element arriving as non-late MUST NOT be changed to late
b) element arriving
Hi Reuven,
To be clear, we already have this solved for BigQueryIO. I am hoping there is a
similar solution for JdbcIO.
Regards,
Willem
On Sun, May 31, 2020, at 12:42 PM, Reuven Lax wrote:
> This should be possible using the Beam programmatic API. You can pass
> BigQueryIO a function that
