[jira] [Commented] (SYSTEMML-529) Maintain column name data when read csv with headers

["Shirish Tatikonda (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SYSTEMML-529?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15158545#comment-15158545 ] 

Shirish Tatikonda commented on SYSTEMML-529:
--------------------------------------------

I think what you are suggesting here comes under the broader topic of creating support for data frames or tables. 
We started in this direction when we introduced {{transform()}} functionality -- the input to this function must have the data type as {{frame}}. However, the design is half-baked. 

Creating the proper support for {{frame}} data type along with well-defined operators is a massive task that requires careful design. We should have a separate JIRA for this.

> Maintain column name data when read csv with headers
> ----------------------------------------------------
>
>                 Key: SYSTEMML-529
>                 URL: https://issues.apache.org/jira/browse/SYSTEMML-529
>             Project: SystemML
>          Issue Type: Improvement
>          Components: APIs
>            Reporter: Deron Eriksson
>
> Column names are a useful form of metadata that describe the data in the columns of a CSV file. This information can be very useful when outputting information about the data, for example when printing out to the console.
> When reading from a CSV file with a header to a matrix variable, it would be nice if the matrix object contained the column names.
> For example, for a read such as the following:
> {code}
> m = read("m.csv", format="csv", header=TRUE);
> {code}
> It would be nice if the user could obtain the column names with something such as:
> {code}
> c = colnames(m);
> print("column names: " + c);
> {code}
> and see something such as:
> {code}
> column names: location, gender, age
> {code}



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