[datatable-help] Using := in .onLoad

Arunkumar Srinivasan aragorn168b at gmail.com
Tue Jan 20 14:32:24 CET 2015


On Tue, Jan 20, 2015 at 2:22 PM, Juan Manuel Truppia <jmtruppia at gmail.com>
wrote:

> Let's follow an example. I'm developing my own package. I add data.table
> to imports in my description file. This means I have to use the :: operator
> each time I want to call a data.table function.  Can I use := in this setup?
>
"I have to use / Can I use" - here you refer to the user using your package
or yourself as the developer trying to use data.table functions and := from
within your package?


>
> On Tue, Jan 20, 2015, 10:01 Arunkumar Srinivasan <aragorn168b at gmail.com>
> wrote:
>
>> Juan,
>>
>>>
>>> I already know that.
>>>
>> Okay, great.
>>
>>> I usually struggle when using data.table inside my own packages. I start
>>> by listing it in imports only (and not importing anything)
>>>
>> "imports only (and not importing anything)"?  - what do you mean?
>>
>>> but end up depending on it, as for the users to be able to manipulate
>>> the data.tables with [ and :=.
>>>
>> It depends on whether you want the data.table NAMESPACE to be attached or
>> not. I think these posts may help:
>>
>> http://r.789695.n4.nabble.com/Re-R-CMD-check-checking-in-development-version-of-R-td4696125.html#none
>>
>> http://stackoverflow.com/questions/8637993/better-explanation-of-when-to-use-imports-depends
>>
>>
>>> I end up importing some or all of data.table also, because I don't know
>>> how to call := the :: notation.
>>>
>> I don't really follow this. Why do you want to use `::` along with `:=`.
>> The last post we just discussed that it can't be done, as it is not
>> designed to be used outside of  `[...]` and in very specific ways. Check
>> data.table:::`:=`. It is designed to error. This post might help:
>> http://stackoverflow.com/q/7033106/559784
>>
>>> My specific problem was on using := during on.Load.
>>>
>> It'd be useful to know what you are trying to do, along with your code.
>>
>>> I already know how to use it, and know that it can only be used inside [.
>>> Maybe it is imperative to @import data.table to be able to use :=?
>>>
>> I don't follow exactly how you are trying to use := to answer this.
>>
>> Arun
>>
>>>
>>> On Mon, Jan 19, 2015, 19:35 Arunkumar Srinivasan <aragorn168b at gmail.com>
>>> wrote:
>>>
>>>> Within the square brackets in "DT[ .... ]" and even there, only in `j`
>>>> (as the error message points out).
>>>>
>>>> On Mon, Jan 19, 2015 at 11:31 PM, Juan Manuel Truppia <
>>>> jmtruppia at gmail.com> wrote:
>>>>
>>>>> Arun, what does "within the frame of data.table" mean?
>>>>>
>>>>> On Mon, Jan 19, 2015, 19:17 Arunkumar Srinivasan <
>>>>> aragorn168b at gmail.com> wrote:
>>>>>
>>>>>> Juan,
>>>>>>
>>>>>> `:=` is designed to be used only within the frame of data.table.
>>>>>>
>>>>>> On Mon, Jan 19, 2015 at 10:45 PM, Juan Manuel Truppia <
>>>>>> jmtruppia at gmail.com> wrote:
>>>>>>
>>>>>>> Hi, I'm using data.table inside my own packages. I'm having some
>>>>>>> trouble using data.table functions on my .onLoad method.
>>>>>>> I'm actually depending (not importing) data.table, but even then, I
>>>>>>> can't get := to work in .onLoad.
>>>>>>> I'm using the `:=`(a = x, b = y) version, and getting
>>>>>>>
>>>>>>> Error in `:=`(a = x) :
>>>>>>> Check that is.data.table(DT) == TRUE. Otherwise, := and `:=`(...)
>>>>>>> are defined for use in j, once only and in particular ways. See help(":=").
>>>>>>>
>>>>>>> If I comment that, and then run it from the console after the
>>>>>>> package is loaded, it works.
>>>>>>>
>>>>>>> Any ideas?
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> datatable-help mailing list
>>>>>>> datatable-help at lists.r-forge.r-project.org
>>>>>>>
>>>>>>> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help
>>>>>>>
>>>>>>
>>>>>>
>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/datatable-help/attachments/20150120/d769c1bf/attachment.html>


More information about the datatable-help mailing list