You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@systemml.apache.org by "Deron Eriksson (JIRA)" <ji...@apache.org> on 2016/02/17 22:14:18 UTC

[jira] [Created] (SYSTEMML-528) Output meaningful column names when write csv with header=TRUE

Deron Eriksson created SYSTEMML-528:
---------------------------------------

             Summary: Output meaningful column names when write csv with header=TRUE
                 Key: SYSTEMML-528
                 URL: https://issues.apache.org/jira/browse/SYSTEMML-528
             Project: SystemML
          Issue Type: Improvement
          Components: APIs
            Reporter: Deron Eriksson
            Priority: Minor


Currently if the write function is called with csv format and header is set to TRUE, column headers are output in [C1,C2,C3...] form, but these column names are not very useful.

For example:
{code}
write(b, "n.csv", format="csv", header=TRUE);
{code}
generates results similar to:
{code}
C1,C2,C3
1.0,2.0,3.0
4.0,0,6.0
7.0,8.0,9.0
{code}

Column names are a form of metadata that describe the data in the file. Therefore, allowing the writing of meaningful column names can assist in the understanding of the data contained in the data file.

A solution involving a matrix of strings would probably yield a relatively clean solution, which could involve a "column-names" parameter that takes a variable that is a 1-row matrix of strings.

Another solution could involve being able to write a "column-names" parameter that took a string value, which would be more of a hard-coded approach, but it wouldn't rely on the implementation of a matrix of strings.





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