Re: [Rd] unique.matrix issue [Was: Anomaly with unique and match]
On Thu, Mar 10, 2011 at 01:19:48AM -0800, Henrik Bengtsson wrote:
It should be possible to run unique()/duplicated() column by column
and incrementally update the set of unique/duplicated rows. This
would avoid any coercing. The benefit should be even greater for
data.frame():s.
This is a good point. An implementation of this using sorting can
be done as follows
Sort the data frame using function order().
Determine the groups of consecutive equal rows in the sorted df.
Map the first row of each group to the original order of the rows.
Since sorting by the function order() is stable, we obtain the first
in each group of equal rows also in the original order.
The coercion approach uses hashing for string comparison, but the
efficiency of hashing seems to be overweighted by the inefficiency
of the coercion. So, we get the following comparison.
a - matrix(sample(c(1234, 5678), 12*1, replace=TRUE), ncol=12)
df - data.frame(a)
do.unique.sort - function(df)
{
i - do.call(order, df)
n - nrow(df)
u - c(TRUE, rowSums(df[i[2:n], ] == df[i[1:(n-1)], ]) ncol(df))
df[u[order(i)], ]
}
system.time(out1 - do.unique.sort(df))
system.time(out2 - unique(df))
identical(out1, out2)
The result may be, for example
user system elapsed
0.279 0.000 0.273
user system elapsed
0.514 0.000 0.468
[1] TRUE
On another computer
user system elapsed
0.058 0.000 0.058
user system elapsed
0.187 0.000 0.188
[1] TRUE
On Thu, Mar 10, 2011 at 01:39:56PM -0600, Terry Therneau wrote:
Simon pointed out that the issue I observed was due to internal
behaviour of unique.matrix.
I had looked carefully at the manual pages before posting the question
and this was not mentioned. Perhaps an addition could be made?
According to the description of unique(), the user may expect that if
b is obtained using
b - unique(a)
then for every i there is j, such that
all(a[i, ] == b[j, ])
This is usually true, but not always, because among several numbers
in a with the same as.character() only one remains in b. If this
is intended, then i support the suggestion to include a note in the
documentation.
Let me add an argument against using as.character() to determine,
whether two numbers are close. The maximum relative difference between
the numbers, which have the same 15 digit decimal representation, varies
by a factor up to 10 in different ranges. Due to this, we have
x - 1 + c(1.1, 1.3, 1.7, 1.9)*1e-14
unique(as.character(x))
[1] 1.01 1.02
unique(as.character(9*x))
[1] 9.1 9.12 9.15
9.17
The relative differences between components of 9*x are the same as the
relative differences in x, but if the mantissa begins with 9, then
a smaller relative difference is sufficient to change 15-th digit.
In terms of unique(), this implies
nrow(unique(cbind(x)))
[1] 2
nrow(unique(cbind(9*x)))
[1] 4
Petr Savicky.
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
[Rd] Run script automatically when package is loaded
Dear list, is it possible to specify a script that is executed automatically when my package is mounted via 'require(my.pkg)' or 'library(my.pkg)'? Id' like to specify execute a small init function that creates some crucial environment structures. As it's always the first thing to do when using the package, I wanted to hide it from the user so he won't have to think about this step. Can I use the lazy-loading functionality of packages for that (Writing R Extensions, Section 1.1.5 Data in packages, pp. 10)? Thanks for any suggestions, Janko __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] WARNING Undocumented S4 methods 'initialize' - why?
Dear all,
Meanwhile I found my mistake.
I forgot to add class QualTreeSet to 'initialize-methods.Rd'.
However, I am not sure if I should make 'initialize' public at all.
Best regards
Christian
On 3/11/11 9:58 PM, cstrato wrote:
Dear all,
I am just writing the documentation file for S4 class 'QualTreeSet' and
get the following warning with R CMD check:
* checking for missing documentation entries ... WARNING
Undocumented S4 methods:
generic 'initialize' and siglist 'QualTreeSet'
All user-level objects in a package (including S4 classes and methods)
should have documentation entries.
See the chapter 'Writing R documentation files' in manual 'Writing R
Extensions'.
All my S4 classes have a method 'initialize' and R CMD check has never
complained. Thus, do you have any idea why I get suddenly this warning
for method 'initialize'?
Here is the code for 'QualTreeSet':
setClass(QualTreeSet,
representation(qualtype = character,
qualopt = character
),
contains=c(ProcesSet),
prototype(qualtype = rlm,
qualopt = raw
)
)#QualTreeSet
setMethod(initialize, QualTreeSet,
function(.Object,
qualtype = rlm,
qualopt = raw,
...)
{
if (qualtype == ) qualtype - rlm;
if (qualopt == ) qualopt - raw;
.Object - callNextMethod(.Object,
qualtype = qualtype,
qualopt = qualopt,
...);
.Object@qualtype = qualtype;
.Object@qualopt = qualopt;
.Object;
}
)#initialize
However, here is my code for a similar class 'ExprTreeSet' (which is the
class from where I have copied the code):
setClass(ExprTreeSet,
representation(exprtype = character,
normtype = character
),
contains=c(ProcesSet),
prototype(exprtype = none,
normtype = none
)
)#ExprTreeSet
setMethod(initialize, ExprTreeSet,
function(.Object,
exprtype = none,
normtype = none,
...)
{
if (exprtype == ) exprtype - none;
if (normtype == ) normtype - none;
.Object - callNextMethod(.Object,
exprtype = exprtype,
normtype = normtype,
...);
.Object@exprtype = exprtype;
.Object@normtype = normtype;
.Object;
}
)#initialize
In this case R CMD check does not complain, so why does it in the case
of 'QualTreeSet'?
Here is my:
sessionInfo()
R version 2.12.1 (2010-12-16)
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
locale:
[1] C
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] xps_1.11.4
Thank you in advance
Best regards
Christian
_._._._._._._._._._._._._._._._._._
C.h.r.i.s.t.i.a.n S.t.r.a.t.o.w.a
V.i.e.n.n.a A.u.s.t.r.i.a
e.m.a.i.l: cstrato at aon.at
_._._._._._._._._._._._._._._._._._
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] Run script automatically when package is loaded
On 12 March 2011 at 14:06, Janko Thyson wrote: | Dear list, | | is it possible to specify a script that is executed automatically when my | package is mounted via 'require(my.pkg)' or 'library(my.pkg)'? That has been possible all along. See help(.onLoad) if you use a NAMESPACE (as you should) or help(.First.lib) if you don't. Dirk | Id' like to specify execute a small init function that creates some crucial | environment structures. As it's always the first thing to do when using the | package, I wanted to hide it from the user so he won't have to think about | this step. | | Can I use the lazy-loading functionality of packages for that (Writing R | Extensions, Section 1.1.5 Data in packages, pp. 10)? | | Thanks for any suggestions, | Janko | | __ | R-devel@r-project.org mailing list | https://stat.ethz.ch/mailman/listinfo/r-devel -- Dirk Eddelbuettel | e...@debian.org | http://dirk.eddelbuettel.com __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
[Rd] par(ask=TRUE) in R CMD check?
Hello: What happens in the auto-checks on R-Forge and CRAN with code using par(ask=TRUE)? Is this routine, or can it create problems? The fda package uses ask=TRUE to provide the user with a way to examine a group of plots. In the past, I've marked those tests in \examples with \dontrun. However, I wonder if that is necessary. I tried it on Windows using R 2.12.0 and R Tools from that version, and R Tools seemed to supply all the mouse clicks required. However, before I SVN Commit to R-Forge, I felt a need to ask. Thanks, Spencer __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] par(ask=TRUE) in R CMD check?
On Sat, 12 Mar 2011, Spencer Graves wrote: Hello: What happens in the auto-checks on R-Forge and CRAN with code using par(ask=TRUE)? Is this routine, or can it create problems? The fda package uses ask=TRUE to provide the user with a way to examine a group of plots. In the past, I've marked those tests in \examples with \dontrun. However, I wonder if that is necessary. I tried it on Windows using R 2.12.0 and R Tools from that version, and R Tools seemed to supply all the mouse clicks required. However, before I SVN Commit to R-Forge, I felt a need to ask. No mouse clicks are needed! par(ask=TRUE) is ignored on a pdf device (which is the non-interactive default device). In any case, the help says ‘ask’ logical. If ‘TRUE’ (and the R session is interactive) the ... This not really a graphics parameter, and its use is deprecated in favour of ‘devAskNewPage’. See also the 'ask' argument in ?example. What you say you want to do is the default behaviour of example(), and has been for some years. -- Brian D. Ripley, rip...@stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595__ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
[Rd] Excited about the near future...
Some already know, but I think it deserves a bit of a attention here as well: It looks like we're about to get new features in R that will be very powerful! That should be a good enough teaser for now... /Henrik PS ...and thanks for making it available plus credits to similar efforts by others. __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
