[R-sig-ME] Fwd: Re: R-core post from jw1085 at wildcats.unh.edu requires approval

[Ben Bolker]

It doesn't look like r-sig-mixed-models was in the Cc: , please keep that in ...

   The list doesn't take attachments, so if you want to include data
either (if it's short) use dput() and cut-and-paste the text, or (if
it's long) post it somewhere.

  In any case, cutting-and-pasting the results of sessionInfo() will
be extremely useful in helping us get farther.  It would also help if
you responded to the question about whether lme4 is a feasible
alternative ...


On Wed, May 10, 2017 at 2:07 PM, Jochen Wirsing <jw1085 at wildcats.unh.edu> wrote:
> Dear all,
>
> thank you very much for you good and helpful input. I would like to
> provide a subset of my data, to show the problem, but don't know whether
> I could (or should) just send it as an attachment.
>
> The code I'm using is:
>
> ### Unconditional Model
> library(nlme)
> library(broom)
>
> uncmod <- lme(fixed = year_c~1, random = ~1|Place, data = data2, method
> ="ML")
> tidy(uncmod)
> summary(uncmod)
> varCorr(uncmod)
>
>
> And the Error I'm getting is
> Error in pdFactor.pdLogChol(X[[i]], ...) : object 'logChol_pd' not found
>
> I'd really appreciate if this issue could be fixed, so that I can keep
> on using R productively in my studies.
>
> Best,
> Jochen Wirsing
>
>
> On 05/08/2017 03:43 PM, Ben Bolker wrote:
>> My only wild guess about this is that there are (still unresolved as
>> far as I know) issues with the latest version of Matrix on 32-bit
>> operating systems (unlikely unless you're using Linux, and even then
>> slightly rare).  If so (sessionInfo() will tell us this), reverting to
>> the previous version of Matrix (e.g. see ?devtools::install_version)
>> should help.
>>
>> On Mon, May 8, 2017 at 2:42 PM, Kevin Wright <kw.stat at gmail.com> wrote:
>>> Somebody might be able to look at your error message and figure out what is
>>> wrong, but often that is not possible just by looking.  If you want people
>>> to help you, then you need to help people understand your problem, perhaps
>>> using simulated data.  See:
>>> http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example
>>>
>>> Kevin
>>>
>>> On Mon, May 8, 2017 at 11:50 AM, Jochen Wirsing <jw1085 at wildcats.unh.edu>
>>> wrote:
>>>
>>>> Dear R Mixed Models SIG,
>>>>
>>>> below you will find an email that I originally sent to the R core group,
>>>> who technically is the maintainer of the nlme package, with which I've
>>>> had some serious issues since the last update. Since this seems to keep
>>>> preventing me from using R for my project, I am reaching out to you,
>>>> hoping this problem can be fixed in a timely manner, as it probably
>>>> affects a lot of other people as well.
>>>>
>>>>
>>>> Thank you very much,
>>>>
>>>> Jochen Wirsing
>>>>
>>>> -------- Forwarded Message --------
>>>> Subject:        Re: R-core post from jw1085 at wildcats.unh.edu requires
>>>> approval
>>>> Date:   Wed, 3 May 2017 15:11:09 +0530
>>>> From:   Deepayan Sarkar <deepayan.sarkar at gmail.com>
>>>> To:     r-core-owner at r-project.org, Jochen Wirsing <
>>>> jw1085 at wildcats.unh.edu>
>>>>
>>>>
>>>>
>>>> Dear Jochen Wirsing,
>>>>
>>>> Although r-core is technically the maintainer of nlme, you should
>>>> write to one of the public R mailing lists get help regarding such
>>>> problems: see
>>>>
>>>> https://www.r-project.org/mail.html
>>>>
>>>> Either R-help or R-SIG-mixed-models should be appropriate.
>>>>
>>>> Best,
>>>> -Deepayan
>>>>
>>>>> ---------- Forwarded message ----------
>>>>> From: Jochen Wirsing <jw1085 at wildcats.unh.edu>
>>>>> To: R-core at r-project.org
>>>>> Cc:
>>>>> Bcc:
>>>>> Date: Mon, 1 May 2017 11:38:09 -0400
>>>>> Subject: Problem with nlme after update to R 3.4.0
>>>>>
>>>>> Dear Sir or Madam,
>>>>>
>>>>> I am a grad student at UNH, teaching myself how to use R. Not having a
>>>> background in programming, it is a bit hard to figure out what's wrong, so
>>>> I thought I should let you know.
>>>>> Last night, I ran an MLM, and it worked just fine. This morning though,
>>>> I saw that there is a new Version of RStudio as well as R itself, so I
>>>> updated both. After doing so, the very identical code that worked last
>>>> night, doesn't work anymore and throws a rather obscure (at least to me)
>>>> error:
>>>>>
>>>>> 15.
>>>>> pdFactor.pdLogChol(X[[i]], ...)
>>>>> 14.
>>>>> FUN(X[[i]], ...)
>>>>> 13.
>>>>> lapply(object, pdFactor)
>>>>> 12.
>>>>> unlist(lapply(object, pdFactor))
>>>>> 11.
>>>>> pdFactor.reStruct(object)
>>>>> 10.
>>>>> pdFactor(object)
>>>>> 9.
>>>>> unlist(pdFactor(object))
>>>>> 8.
>>>>> MEEM(object, conLin, control$niterEM)
>>>>> 7.
>>>>> Initialize.reStruct(X[[i]], ...)
>>>>> 6.
>>>>> FUN(X[[i]], ...)
>>>>> 5.
>>>>> lapply(object, Initialize, data, conLin, control)
>>>>> 4.
>>>>> Initialize.lmeStruct(lmeSt, dataMix, grps, control = controlvals)
>>>>> 3.
>>>>> Initialize(lmeSt, dataMix, grps, control = controlvals)
>>>>> 2.
>>>>> lme.formula(fixed = year_c ~ 1, random = ~1 | ID, data = data, method =
>>>> "ML")
>>>>> 1.
>>>>> lme(fixed = year_c ~ 1, random = ~1 | ID, data = data, method = "ML")
>>>>>
>>>>>
>>>>>
>>>>> The code I used was:
>>>>>
>>>>> ### Unconditional Model
>>>>> library(nlme)
>>>>> library(broom)
>>>>>
>>>>> uncmod <- lme(fixed = year_c~1, random = ~1|ID, data = data, method
>>>> ="ML")
>>>>> tidy(uncmod)
>>>>> summary(uncmod)
>>>>> varCorr(uncmod)
>>>>>
>>>>>
>>>>>
>>>>> I hope this helps and the problem can be fixed soon, because right now,
>>>> I am stopped dead in my work, unable using the nlme package, on which my
>>>> analysis depends.
>>>>>
>>>>> Thank you very much for your consideration, help, and good work!
>>>>>
>>>>>
>>>>> Best,
>>>>>
>>>>>
>>>>> Jochen Wirsing
>>>>
>>>>         [[alternative HTML version deleted]]
>>>>
>>>> _______________________________________________
>>>> R-sig-mixed-models at r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>>>>
>>>
>>>
>>> --
>>> Kevin Wright
>>>
>>>         [[alternative HTML version deleted]]
>>>
>>> _______________________________________________
>>> R-sig-mixed-models at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>