Re: Industrializing batch ML algorithm using Apache Beam/Dataflow (on Google Cloud Platform)

Sorry Germain ! and thanks again :) Le mar. 16 juil. 2019 à 15:00, Massy Bourennani a écrit : > Hi David ! > this helps a lot, > Many thanks :) > Massy > > Le mar. 16 juil. 2019 à 11:12, Germain Tanguy < > germain.tan...@dailymotion.com> a écrit : > >> Hello

Re: Industrializing batch ML algorithm using Apache Beam/Dataflow (on Google Cloud Platform)

ataflow code to have the same > syntax as sklearn. Then, DS will be able to be autonomous with the > scalability and without to know the complexity of the cluster-computing > framework. > > > > Hope this helps. > > > > Germain. > > > > *From: *Massy Boure

Industrializing batch ML algorithm using Apache Beam/Dataflow (on Google Cloud Platform)

Hi all, Here is the link to the Reddit post[1] Many thanks for your help. Massy [1] https://www.reddit.com/r/dataengineering/comments/cdp5i3/industrializing_batch_ml_algorithm_using_apache/

Re: How Can I access the key in subclass of CombinerFn when combining a PCollection of KV pairs?

I was looking for some built-in feature, as suggested in the programming guide. However the answer does the job. Massy Le ven. 7 juin 2019 à 22:13, Kenneth Knowles a écrit : > The answer on StackOverflow looks good to me. > > Kenn > > On Fri, Jun 7, 2019 at 4:11 AM Massy Bour

How Can I access the key in subclass of CombinerFn when combining a PCollection of KV pairs?

Hi, here is the SO post: https://stackoverflow.com/questions/56451796/how-can-i-access-the-key-in-subclass-of-combinerfn-when-combining-a-pcollection Many thanks

Why Apache beam can't infer the default coder when using KV?

Hi, I'm looking for some help on Apache Airflow. Here is the SO post: https://stackoverflow.com/questions/56443416/why-apache-beam-cant-infer-the-default-coder-when-using-kvstring-string Many thanks, Massy