gmane.comp.lang.r.general

Re: Query about creating time sequences

2012-05-26T00:52:24

One (somewhat kludgy) way would be to use seq() to make one day's worth of times then to pass those to outer() to add in the needed days and then coerce the whole thing back to a sorted vector. 

I'm not at a computer right now so this won't be quite right but something like


x <- seq(x.start.first.day, x.end.first.day, by = "sec")

y <- 24*60*60 *(1:n.days)

sort(as.vector(outer(x, y, "+")))

Changing the order of x and y might make the sort unnecessary. 

M

On May 25, 2012, at 1:14 PM, Shivam <shivamsingh< at >gmail.com> wrote:

Re: How to change width of bar when there are very few bars?

2012-05-26T00:45:19

You can supply a width argument but you also need xlim if I remember right for the specification to have visible effect. 

Michael

On May 25, 2012, at 2:03 PM, Manish Gupta <mandecent.gupta< at >gmail.com> wrote:

Re: evaluate whether function returns error

2012-05-26T00:43:53

I'm not sure I follow... It's an error -- what do you mean by "evaluate"?

If you are looking to catch errors, you can do so with try or tryCatch and then check for inherits(x, "try-error") to see if there was an error. See the examples for details of how exactly to set this up -- I'm not at a computer right now and my memory is fuzzy. 

M

On May 25, 2012, at 4:52 PM, mdvaan <mathijsdevaan< at >gmail.com> wrote:

Re: Multiple rms summary plots in a single device

Frank Harrell — 2012-05-25T23:50:17

Mike I don't know what you are trying to do with plot(summary( ... ~ ...)) as
this is not calling an rms function such as summary.rms.

To use mfrow and to control which variables are placed on one panel using
plot(summary(rms model fit object)) specify est.all=FALSE to summary( )
after specifying the limited list of variables to plot in the current panel.

Frank

Mike Harwood-2 wrote


-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context: http://r.789695.n4.nabble.com/Multiple-rms-summary-plots-in-a-single-device-tp4631330p4631423.html
Sent from the R help mailing list archive at Nabble.com.

Re: applying cbind (or any function) across all components in a list

Rui Barradas — 2012-05-25T22:43:07

Hello,

The use of function(i) is because there are two list to be processed, 
with just one I would have used a simpler form of lapply. I don't 
rebember exactly who wrote this in a post some time ago (Michael 
Weylandt?) but imagine a list is a train. Then, list[ i ] is a car and 
list[[ i ]] is the passengers. In your case they are matrices. In the 
case only one list was to be processed, this would do:

lapply(l1, colSums) # for each matrix, apply the function
# more complicated function, typically, not part of base R (or other 
packages)
lapply( l1, function(x) log(1 + colSums(x)) )
# same as above, but the former uses an unnamed, temporary, function, 
that we can dispose of.
f <- function(x) log(1 + colSums(x))
lapply(l1, f)

In my previous post, seq_len guarantees that the index vector is well 
formed. If the list has zero elements, the form 1:length(l1) becomes 1:0 
== c(1, 0) but seq_len(0) == integer(0).
Then, function(i) is a function of the indices 1, 2, ..., length(l1). 
And both 'l1' and 'l2' c

Re: sweave tables as images?

Yihui Xie — 2012-05-25T22:36:20

The hook function hook_pdfcrop() also works on PNG images, although
the function name is somehow misleading. The requirement is that you
have ImageMagick installed and the command 'convert' is in your PATH.
It is a contribution from < at >r2d3
https://github.com/yihui/knitr/issues/209#issuecomment-5266580

Regards,
Yihui
--
Yihui Xie <xieyihui< at >gmail.com>
Phone: 515-294-2465 Web: http://yihui.name
Department of Statistics, Iowa State University
2215 Snedecor Hall, Ames, IA


On Fri, May 25, 2012 at 4:16 PM, Alexander Shenkin <ashenkin< at >ufl.edu> wrote:

"shift" + "arrow" in getGraphicsEvent

cgenolin — 2012-05-25T22:33:40

Hi the list,

Do you know if getGraphicsEvent can detect the use of keys combinaison like
"shift+arrow" or "Crlt+arrow" ?

Best
Christophe

--
View this message in context: http://r.789695.n4.nabble.com/shift-arrow-in-getGraphicsEvent-tp4631419.html
Sent from the R help mailing list archive at Nabble.com.

Re: sweave tables as images?

Yihui Xie — 2012-05-25T22:32:04

Normally I hate the device()+code+dev.off()+cat('\\includegraphics{})
trick. Let me show you how trivial this can be in knitr:
https://gist.github.com/2790922

This example shows the real source code to the reader
(device()+dev.off() makes no sense to them), and leaves all the dirty
tricks behind the scene.

That being said, Sweave can be regarded as a subset of knitr, so if
you really want to play the old trick, you still can do it in knitr.

Regards,
Yihui
--
Yihui Xie <xieyihui< at >gmail.com>
Phone: 515-294-2465 Web: http://yihui.name
Department of Statistics, Iowa State University
2215 Snedecor Hall, Ames, IA


On Fri, May 25, 2012 at 5:01 PM, baptiste auguie
<baptiste.auguie< at >googlemail.com> wrote:

Re: How to remove square brackets, etc. from address strings?

Sarah Goslee — 2012-05-25T22:19:04

Part of your problem is that your regexes have spaces in them, so
that's what you're matching.

A small reproducible example would be more useful. I'm not feeling
inclined to wade through all your linked files on Friday evening, but
see if this helps:

[1] "New Zealand" "USA"         "Germany"     "Germany"     "Germany"
   "Germany"     "Germany"     "Germany"


Sarah

On Fri, May 25, 2012 at 4:31 PM, Sabina Arndt <sabina.arndt< at >hotmail.de> wrote:

Re: sweave tables as images?

baptiste auguie — 2012-05-25T22:01:49

My guess is that it would be impossible to use the table size in
Sweave, only mildly non-trivial in knitr with a proper hook. The
easiest hack in either of them would be to manually open the png
device, and use cat() with \includegraphics{} where you want the plot
diplayed.

HTH,

b.

On 26 May 2012 09:43, Alexander Shenkin <ashenkin< at >ufl.edu> wrote:

Re: sweave tables as images?

Alexander Shenkin — 2012-05-25T21:43:37

this works - thanks baptiste!  i'm working in Sweave right now - perhaps
it will be tough in knitr as you mention.

On 5/25/2012 4:31 PM, baptiste auguie wrote:

Re: sweave tables as images?

baptiste auguie — 2012-05-25T21:31:51

you can open a device that has the exact dimensions of the table,

g = tableGrob(head(iris, 4))
png("test.png", width=convertWidth(grobWidth(g), "in", value=TRUE),
        height=convertHeight(grobHeight(g), "in",
value=TRUE),units="in", res=150)
grid.draw(g)
dev.off()

Doing this with knitr might be tricky though, since the unit
conversion opens a blank device window, and you'd want to define some
hook instead of manually creating the png file.

I have another version of grid.table where you can specify the width
and height manually [*], e.g to span the full window, but it's not
necessarily a desirable thing (the spacing between rows and columns
can become too large).


HTH,

baptiste


[*] experimental code at https://gist.github.com/2013903

On 26 May 2012 09:16, Alexander Shenkin <ashenkin< at >ufl.edu> wrote:

Re: applying cbind (or any function) across all components in alist

Hans Thompson — 2012-05-25T21:30:23

Thank you everyone for working through the confusion from me posting from
Nabble and missing context.

Both Rui's and David's solutions are working for my problem.  Rui's first
interpretation is the application I was looking for but I was also more
generally interested in how to do the second one if I wanted.

I still don't understand how I would use function(i) for working with
components of lists in the future though.  Is there a simpler example?  I'm
going to play with using seq_len also.

On Fri, May 25, 2012 at 8:20 AM, Rui Barradas <ruipbarradas< at >sapo.pt> wrote:


[[alternative HTML version deleted]]

Re: sweave tables as images?

Alexander Shenkin — 2012-05-25T21:16:33

Thanks Yihui,

That's a great idea, and comes close to the mark, except that I have to
use png's in order to "Insert & Link" them as pictures in Word (and
hence make the doc both shareable and update when new figures are
generated).

thanks,
allie

On 5/25/2012 2:57 PM, Yihui Xie wrote:

Re: question about TryCatch and lapply

Martin Morgan — 2012-05-25T20:57:58

In the context of tryCatch in your question

   lst <- list(1, 2, -3, 4)
   sapply(lst, function(x) tryCatch({
       stopifnot(x > 0)
       x
   }, error=function(err) {
       NA
   }))

which separates out the implementation of the function from the error 
handling instead of intermingling the two as with try(). Or a little 
more elaboration, reporting the error to stderr as might be appropriate 
for a parallel job

   sapply(lst, function(x) {
       tryCatch({
           stopifnot(x>0)
           sqrt(x)
       }, error=function(err) {
           ## demote to message on stderr; in context of tryCatch so
           ## 'x' visible
           message(conditionMessage(err), "; x = ", x)
           NaN
       })
   })

or for warnings, showing use of restarts

   withCallingHandlers({
       sapply(lst, sqrt)
   }, warning=function(warn) {
       message(conditionMessage(warn))
       ## restart where the warning occurs; see ?warning and source for
       ## warning()
       invokeRestart("muffleWarning")

Re: Could "incomplete final line found" be more serious than a warning?

Michael Bärtl — 2012-05-25T20:45:15

Dear all,

thank you very much for the prompt reply by Professor Dalgaard!
Unfortunately, the option he recommended doesn't work for me but still 
gives me the same error.
But when I manually change the encoding of the text file and save it, my 
script works properly again.
So, thanks a lot again! I hadn't thought the encoding could be to blame 
in this case. Now I'm just glad this is solved!
Best wishes,

Michael Baertl

evaluate whether function returns error

mdvaan — 2012-05-25T20:52:35

Hi,

The following returns an error message. How do I evaluate (TRUE or FALSE)
the function?

require(XML)
readHTMLTable("http://www.sec.gov/Archives/edgar/data/2969/000095012399010952/0000950123-99-010952.txt")

Thanks in advance!

Math

--
View this message in context: http://r.789695.n4.nabble.com/evaluate-whether-function-returns-error-tp4631406.html
Sent from the R help mailing list archive at Nabble.com.

change colors/ symbols of lda plots

2012-05-25T19:50:11

Greetings R experts,

I am running a simple lda on some simulation data of mine to show an illustration of my multivariate response data, since it is a simulation I have a very large amount of data and the default on plot seems to plot the category names. This is very difficult to interpret even changing the abbreviations. At the expense of sounding naive, my question(s) are:

How can I color code my four categories?

and how can I change the unit display to something simple like a dot, and not the names or abbreviated names of the categories?


Cheers,
-Dan

[[alternative HTML version deleted]]

Selecting with mouse the lines drawn by matplot()

servet cizmeli — 2012-05-25T20:52:43

Dear all,

I have a dataset of spectral measurements of sunlight. One row for each 
different observation, one column for each spectral channel.

I would like to select the spectral curves drawn with matplot() (I only 
know matplot() as a way of easily drawing them) with the help of the 
mouse and extract their row indexes  :

a=url("http://ekumen.homelinux.net/spectra.RData")
load(a)
matplot(lbd2, t(mySpectra), type="l",xlab="Wavelength [nm]",ylab="Radiance")

I was hoping that identify() would work on such a graphic but 
unfortunately it does not. I tried to the "identify" function of the 
package playwith(), but still the same problem.

I think I will have to write a function that does the job. But before 
that, I would like to have your opinions. What would be the easiest way 
of accomplishing this task?

Thank you very much
Best regards
servet

Re: subset columns from list with variable substitution

MacQueen, Don — 2012-05-25T20:43:57

The select argument to subset() is supposed to name the columns you want
to keep.
So is the syntax I gave, table[,list1], and it is the correct way when
list1 is a character vector (which it is).

Your error message says that at least one of the values in list1 is not
the name of a column in your data frame.


So, compare
  names(table)
with
   list1
(it looks like your data frame doesn't really have the column names you
think it has)

If you just type
  i %in% namelist
by itself, what do you get? Is it what you are trying to loop over?
Heck, what do you get from
  print(i)
?

If you are trying to loop over the names in namelist, you MUST use
  for (i in namelist)
the expression  i %in% namelist
will give you some combination of logical values TRUE and/or FALSE,
depending on the value of i at the time when you use the expression.

So if it "doesn't work", it's for some other reason. Also, just saying it
doesn't work isn't enough information. Please read and follow the posting
guide (mentioned at the end of e

Re: Reading a bunch of csv files into R

Rich Shepard — 2012-05-25T20:34:36


   I'd put the list of commands in a shell script; e.g.,

# import_files.R
read_csv("file1.csv", header = T, sep = ',')
read_csv("file2.csv", header = T, sep = ',')

etc.

   Then, within a R session, run

source("/full/path/to/import_files.R")

or, on the command line,

R CMD BATCH /full/path/name/to/import_files.R

   You'll still need to list all the files individually unless you use a
loop. I've not tried that so someone more experienced will advise you on
that.

Rich