[
https://issues.apache.org/jira/browse/MAHOUT-1490?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13948962#comment-13948962
]
Dmitriy Lyubimov commented on MAHOUT-1490:
------------------------------------------
could be, i have no opinion on that; but that's not the primary objective
Scala does not support square brackets for this purpose. See Scala bindings
documentation for slicing. For the most part it is done using overloaded
magic methods (apply, update).
in case of data frames of course indexes of slices are not just natural
numbers anymore, although this set of course should be still supported.
i.e. (5, ::) selects 5th row, (3 until 5, 2 until 7) selects slice etc etc.
There are 2 main differences of data frames from matrices from functional
perspective:
(1) slicing by names in addition to indices. This basically means slicing
by sequences of Strings (or individual name). something along the lines
dframe(::,("A","B","C")) to select columns A, B, C; same for the rows.
(2) slices are not mahout-math vectors since they main contain not just
doubles, but potentially Strings as well. (a few predefined simple types).
Again, reading spark bindings section of the manual will help.
3) Retrieve a data frame row slice with the single square bracket []
As per current maven configuration (of math-scala module). at the moment, i
believe, it is 2.10.3. but it shouldn't matter really. Maven and IntelliJ
should handle this all automatically.
patch or github, yes, would be good.
this should be in the sparkbindings package of spark module.
> Data frame R-like bindings
> --------------------------
>
> Key: MAHOUT-1490
> URL: https://issues.apache.org/jira/browse/MAHOUT-1490
> Project: Mahout
> Issue Type: New Feature
> Reporter: Saikat Kanjilal
> Assignee: Dmitriy Lyubimov
> Original Estimate: 20h
> Remaining Estimate: 20h
>
> Create Data frame R-like bindings for spark
--
This message was sent by Atlassian JIRA
(v6.2#6252)
