[jira] [Commented] (BEAM-264) Matrix, Vector, Tensor Abstractions for Inputs, Outputs and Processing and Construction of Such Structures from Scalars, Other Structures and Queries

["Davor Bonaci (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/BEAM-264?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15275707#comment-15275707 ] 

Davor Bonaci commented on BEAM-264:
-----------------------------------

Can you develop your idea a little bit please?

I'd suggest working through a scenario, trying to express it in the Beam model, and then trying to identify areas for improvement. The things you mention can be expressed in the Beam model already. There are certainly areas for improvement, however. To be able to assess this JIRA issue, we'll need more data as to what actually do you propose.

> Matrix, Vector, Tensor Abstractions for Inputs, Outputs and Processing and Construction of Such Structures from Scalars, Other Structures and Queries
> -----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: BEAM-264
>                 URL: https://issues.apache.org/jira/browse/BEAM-264
>             Project: Beam
>          Issue Type: Improvement
>          Components: beam-model
>            Reporter: Suminda Dharmasena
>            Assignee: Davor Bonaci
>
> Not all timeseries data is scalar values hence support other types of values like vectors, matrices. Also ability to construct these constructs. E.g. if I want a time series vector of asset prices as input to a pipeline. Also I will have to choose what stocks are in the vector and also how to handle if the choosing is time varying.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)