[datatable-help] Generating data.tables from a j function

[Cook, Malcolm]

Result.dt<-dt[ , list(sd.dtable=dtable(.SD)), by=trichotomous.variable]  # untested

will return a Result.dt with 2 columns: trichotomous.variable and sd.dtable

The value of Result.dt[1,sd.dtable][[1]] will give the 1st groups dtable results

~malcolm_cook at stowers.org

From: datatable-help-bounces at lists.r-forge.r-project.org [mailto:datatable-help-bounces at lists.r-forge.r-project.org] On Behalf Of Farrel Buchinsky
Sent: Friday, October 20, 2017 1:14 PM
To: datatable-help at lists.r-forge.r-project.org
Subject: [datatable-help] Generating data.tables from a j function

I am relatively certain that data.table can handle a function in the J position that generates a data.frame or a data.table. However, how does one handle the output? Can you please point me to a url that discusses the capability?

For instance, I sometimes use the package dtables to generate descriptive statistics. It generates a table for

So one could go

library(dtables)
dtable(dt)

But what if one wanted to get descriptive statistics by a trichotomous variable

dt[ , dtable(.SD), by=trichotomous.variable]

I am looking for a general article or blogpost about how to handle outputs that are not single numbers or two numbers. There are functions that yield data.tables or lists or specialized objects such a models. Any suggestions?

I know one can simply print the output to the console

dt[ , print(dtable(.SD)_, by=trichotomous.variable]

But that does not let me do anything intelligent afterwards.
--
Farrel Buchinsky
(412) 567-7870 (gets me everywhere)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/datatable-help/attachments/20171020/1795c15f/attachment.html>