Re: [Rd] uniform sampling without replacement algorithm

Splus used a similar method for sampling from "bigdata" objects. One problem was that sample() is used both for creating a sample and for scrambling the order of a vector. Scrambling the order of a big vector wastes time. It would be nice to be able to tell sample() that we don't care about the

Re: [Rd] file.copy(from=Directory, to=File) oddity

Bug 17337. Note that I get R making the zero-length file on both Windows and Linux, but the return values are different. Bill Dunlap TIBCO Software wdunlap tibco.com On Mon, Sep 11, 2017 at 7:01 AM, Martin Maechler <maech...@stat.math.ethz.ch > wrote: > >>>>> Willi

[Rd] file.copy(from=Directory, to=File) oddity

When I mistakenly use file.copy() with a directory for the 'from' argument and a non-directory for the 'to' and overwrite=TRUE, file.copy returns FALSE, meaning it could not do the copying. However, it also replaces the 'to' file with a zero-length file. dir.create( fromDir <- tempfile() )

[Rd] arithmetic with zero-column data.frames

Should arithmetic operations work on zero-column data.frames (returning a zero-column data.frame with the same number of rows as the data.frame argument(s))? Currently we get: > 1 + data.frame(row.names=c("A","B")) Error in data.frame(value, row.names = rn, check.names = FALSE, check.rows =

Re: [Rd] force promises inside lapply

quote(expr) will make no changes in expr, it just returns its one argument, unevaluated. substitute could be used in your lapply(..., library) example to give library a name instead of a character string for an input (which might be necessary if the character.only argument were not available)

Re: [Rd] force promises inside lapply

Functions, like your loader(), that use substitute to let users confound things and their names, should give the user a way to avoid the use of substitute. E.g., library() has the 'character.only' argument; if TRUE then the package argument is treated as an ordinary argument and not passed

Re: [Rd] force promises inside lapply

1: substitute(), when given an argument to a function (which will be a promise) gives you the unevaluated expression given as the argument: > L <- list(a=1, b=2, c=3) > str(lapply(L, function(x) substitute(x))) List of 3 $ a: language X[[i]] $ b: language X[[i]] $ c: language X[[i]] The 'X'

[Rd] matrices with names

How should R deal with matrices that have a 'names' attribute? S (and S+) did not allow an object to have both dims and names but R does. However, some R functions copy the dims but not the names to the returned value and some copy both. I don't see a pattern to it. Is there a general rule for

Re: [Rd] 'do.call' appears to show inconsistent behavior for arguments of format "pkg::fun"

>The function Rmpi::mpi.bcast.cmd() calls eventually something along the lines of > >> scmd <- scmd <- substitute(cmd) >> arg <- list(...) >> scmd.arg <-serialize(list(scmd=scmd, arg=arg), NULL) >> if (length(scmd.arg$args) > 0) >>do.call(as.character(scmd.arg$scmd), scmd.arg$args, envir =

[Rd] attributes on symbols

The multcomp package has code in multcomp:::expression2coef that attaches the 'coef' attribute to symbols. Since there is only one symbol object in a session with a given name, this means that this attaching has a global effect. Should this be quietly allowed or should there be a warning or an

Re: [Rd] R history: Why 'L; in suffix character ‘L’ for integer constants?

re wdunlap tibco.com On Fri, Jun 16, 2017 at 11:53 AM, peter dalgaard <pda...@gmail.com> wrote: > > Wikipedia claims that C ints are still only guaranteed to be at least 16 bits, and longs are at least 32 bits. So no, R's integers are long. > > -pd > > > On 16 Jun 2017, at

Re: [Rd] R history: Why 'L; in suffix character ‘L’ for integer constants?

But R "integers" are C "ints", as opposed to S "integers", which are C "long ints". (I suppose R never had to run on ancient hardware with 16 bit ints.) Bill Dunlap TIBCO Software wdunlap tibco.com On Fri, Jun 16, 2017 at 10:47 AM, Yihui Xie wrote: > Yeah, that was what I

Re: [Rd] spurious warning in ave()

You can avoid the warnings and the unneeded calls to FUN by adding drop=TRUE to the call to ave(), since all of its ... arguments are passed to interaction (I think). In TERR we dealt with this problem by adding drop=TRUE to ave's argument list and we pass ... and drop=drop to interaction. I'm

Re: [Rd] [WISH / PATCH] possibility to split string literals across multiple lines

If you are changing the parser (which is a major change) you might consider treating strings in the C/C++ way: char *s = "A" "B"; means the same as char *s = "AB"; I am not a big fan of that syntax but it is widely used. A backslash at the end of the line leads to errors

Re: [Rd] stopifnot() does not stop at first non-TRUE argument

While you are fiddling with stopifnot(), please consider changing the form of the error thrown so that it includes the caller's call. The change would be from something like stop( <> ) to stop(simpleError( <>, sys.call(-1))) For the following code f <- function(x, y) { stopifnot(x >

Re: [Rd] R-3.3.3/R-3.4.0 change in sys.call(sys.parent())

gt; wrote: > On Wed, May 10, 2017 at 2:36 AM, William Dunlap via R-devel > <r-devel@r-project.org> wrote: > > Some formula methods for S3 generic functions use the idiom > > returnValue$call <- sys.call(sys.parent()) > > to show how to recreate the returned object or to

Re: [Rd] tempdir() may be deleted during long-running R session

9-4973 > IM: bgpbraverock > > On Wed, 2017-04-26 at 08:25 -0700, William Dunlap via R-devel wrote: > > The Ubuntu machine I use a lot (along with others) must not be > > cleaning > > /tmp as it has a fair number of Rtmp* directories in /tmp, even when > > the

Re: [Rd] tempdir() may be deleted during long-running R session

The Ubuntu machine I use a lot (along with others) must not be cleaning /tmp as it has a fair number of Rtmp* directories in /tmp, even when there are no R sessions running on the machine. I would like to automate their removal but there is no obvious way to see if the R process that created the

[Rd] R_CMethodDef incompatibility (affects R_registerRoutines)

I recently noticed a change between R-3.3.3 and R-3.4.0 in the definition of the R_CMethodDef struct. typedef struct { const char *name; DL_FUNC fun; int numArgs; - R_NativePrimitiveArgType *types; - R_NativeArgStyle *styles; - } R_CMethodDef; I

Re: [Rd] system/system2 and open file descriptors

In S+ on Unix-alikes we dealt with this issue by using fcntl(fd, F_SETFD, 1) to set the close-on-exec flag on a file descriptor as soon as we opened it. Bill Dunlap TIBCO Software wdunlap tibco.com On Wed, Apr 19, 2017 at 8:40 PM, Winston Chang wrote: > In addition to

Re: [Rd] A trap for young players with the lapply() function.

>I think that the suggestion I made, in response to a posting by Barry >>Rowlingson, that the first argument of lapply() be given the name of ".X" >rather >than just-plain-X, would be (a) effective, and (b) harmless. It would break any call to *apply() that used X= to name the first argument.

Re: [Rd] outer not applying a constant function

> Or is this a bad idea? I don't like the proposal. I have seen code like the following (in fact, I have written such code, where I had forgotten a function was not vectorized) where the error would have been discovered much later if outer() didn't catch it. > outer(1:3, 11:13, sum) Error

Re: [Rd] Fwd: Possible memory problems with mallloc when called with .C()

> void dnk_c(double *sortedFsample, unsigned long int n, unsigned long int k, double *dKol) All arguments to C functions called by .C() must be pointers. Also, R integers are C ints, not unsigned long ints. Bill Dunlap TIBCO Software wdunlap tibco.com On Mon, Mar 20, 2017 at 5:55 AM, Hristo

Re: [Rd] Support for user defined unary functions

I can see that allowing a user-defined unary prefix operator can be useful. We want to make sure its precedence and associative behavior are convenient for a variety of envisioned uses, as we won't get a chance to change them after the language construct is introduced. An example of precedence

Re: [Rd] Support for user defined unary functions

OK. I am more concerned now with semantics than the syntax. Bill Dunlap TIBCO Software wdunlap tibco.com On Fri, Mar 17, 2017 at 1:09 PM, Gabriel Becker wrote: > Bill, > > Right. My example was the functional form for clarity. > > There is a desire for a unary-operator

Re: [Rd] Support for user defined unary functions

Your example x = 5 exp = parse(text="f(uq(x)) + y +z") # expression: f(uq(x)) +y + z do_unquote(expr) # -> the language object f(5) + y + z could be done with the following wrapper for bquote my_do_unquote <- function(language, envir = parent.frame()) { if

Re: [Rd] Support for user defined unary functions

>After off list discussions with Jonathan Carrol and with >Michael Lawrence I think it's doable, unambiguous, >and even imo pretty intuitive for an "unquote" operator. For those of us who are not CS/Lisp mavens, what is an "unquote" operator? Can you expression quoting and unquoting in R syntax

Re: [Rd] Support for user defined unary functions

I am biased against introducing new syntax, but if one is experimenting with it one should make sure the precedence feels right. I think the unary and binary minus-sign operators have different precedences so I see no a priori reason to make the unary and binary %xxx% operators to be the same.

[Rd] simplify2array(higher=TRUE, listOfSingleRowDataframes)

I noticed that simplify2array acted oddly when given a list of data.frames of various sizes. If the data.frames have one row, it makes a new first dimension with the dimname equal to rownames(firstDataframe). That dimension does not appear for data.frames with other numbers of rows. >

Re: [Rd] Error in formatDL(nm, txt, indent = max(nchar(nm, "w")) + 3)

It happens in the fda package because some of the headers are longer than typical (e.g., "Repository/R-Forge/DateTimeStamp") and formatDL dies if the indent argument is too large compared to the width argument. It might be nice to change formatDL so it never gave such an error, but did something

Re: [Rd] Error in formatDL(nm, txt, indent = max(nchar(nm, "w")) + 3)

This error can arise when getOption("width") is too small. 80 seems to be the limit for me with R-3.3.2 on Windows. Bill Dunlap TIBCO Software wdunlap tibco.com On Wed, Mar 8, 2017 at 10:28 PM, Spencer Graves wrote: > Hello: > > > I tried "debug(help)" with

Re: [Rd] can we override "if" in R?

Da Zheng would like to override 'if' and 'while' to accept more than scalar logicals and Martin Maechler would like to change 'if' to accept only scalar logicals. No one has mentioned '||' and '&&', which also want scalar logicals. Perhaps a solution is to have all of these call a new generic

Re: [Rd] can we override "if" in R?

dplyr::translate_sql() redefines lots of functions, include "if", to translate from R syntax to SQL syntax. > dplyr::translate_sql(if ("mpg">25) "better" else "worse") CASE WHEN ('mpg' > 25.0) THEN ('better') ELSE ('worse') END Bill Dunlap TIBCO Software wdunlap tibco.com On Sat, Mar 4, 2017

Re: [Rd] Grapics Device Resolution Limits

Were you suppressing warnings? I get a warning along with the "unable to start device 'png'" in some cases where it fails. E.g., on Linux > png("Figure1A.png", h = 7, w = 7, res = 1e5, units = "cm") Error in png("Figure1A.png", h = 7, w = 7, res = 1e+05, units = "cm") : unable to start device

Re: [Rd] issue with unz()?

If you use check.names=FALSE in your call to read.csv you can see that the first column name starts with the 3 bytes ef bb bf, which is the UTF-8 "byte-order mark" that Microsoft applications like to put at the start of a text file stored in UTF-8. > v0514 <- read.csv(unz(temp, file0514[1]),

Re: [Rd] Undefined behavior of head() and tail() with n = 0

In addition, signed zeroes only exist for floating point numbers - the bit patterns for as.integer(0) and as.integer(-0) are identical. Bill Dunlap TIBCO Software wdunlap tibco.com On Thu, Jan 26, 2017 at 1:53 AM, Martin Maechler wrote: >> Florent Angly

Re: [Rd] RFC: tapply(*, ..., init.value = NA)

It would be cool if the default for tapply's init.value could be FUN(X[0]), so it would be 0 for FUN=sum or FUN=length, TRUE for FUN=all, -Inf for FUN=max, etc. But that would take time and would break code for which FUN did not work on length-0 objects. Bill Dunlap TIBCO Software wdunlap

Re: [Rd] How to handle INT8 data

If these are identifiers, store them as strings. If not, what sort of calculations do you plan on doing with them? Bill Dunlap TIBCO Software wdunlap tibco.com On Fri, Jan 20, 2017 at 6:33 AM, Nicolas Paris wrote: > Hello r users, > > I have to deal with int8 data with

[Rd] strsplit(perl=TRUE), gregexpr(perl=TRUE) very slow for long strings

While doing some speed testing I noticed that in R-3.2.3 the perl=TRUE variants of strsplit() and gregexpr() took time proportional to the square of the number of pattern matches in their input strings. E.g., the attached test function times gsub, strsplit, and gregexpr, with perl TRUE (PCRE) and

Re: [Rd] [R-pkg-devel] accessing data by packagename::dataname from within package code fails.

You can define the data in the R directory. You can keep it all in a *.R file by wrapping the text of the *.csv file in quotes and using read.table(text="quoted stuff"), as in: theData <- read.csv(header=TRUE, text=" English,Digit One,1 Two,2 Three,3") N <- nrow(theData) You need to make sure

Re: [Rd] as.formula("x") error on C stack limit

Another example uses formula.character's other arguments: > as.formula("env") Error: object of type 'special' is not subsettable > as.formula("...") Error in eval(expr, envir, enclos) : '...' used in an incorrect context It may happen for the same reason that the following does not give an

Re: [Rd] anonymous function parsing bug?

Here is a simplified version of your problem > { sqrt }(c(2,4,8)) [1] 1.414214 2.00 2.828427 Do you want that to act differently? Bill Dunlap TIBCO Software wdunlap tibco.com On Fri, Oct 21, 2016 at 6:10 AM, Wilm Schumacher wrote: > Hi, > > I hope this is

Re: [Rd] error handling in strcapture

t seems appropriate. > >> > > >> > If there are no matches, then there is no easy way to see if the > >> > prototype > >> > is compatible with the pattern, so should strcapture just assume the > >> > best > >> > and fill in the prototype wit

Re: [Rd] error handling in strcapture

ves NA's when the format does not match the text input. > > > > > > Bill Dunlap > > TIBCO Software > > wdunlap tibco.com > > > > On Wed, Sep 21, 2016 at 2:10 PM, Michael Lawrence > > <lawrence.mich...@gene.com> wrote: > >> > >> Hi Bill,

Re: [Rd] suggested addition to model.matrix

In addition, there is a formula method for data.frame that assumes the first column is the dependent variable. > z <- data.frame(X1=1:6,X2=letters[1:3],Y=log(1:6)) > formula(z) X1 ~ X2 + Y > colnames(model.matrix(formula(z), z)) [1] "(Intercept)" "X2b" "X2c" "Y" Spencer's

Re: [Rd] Undocumented 'use.names' argument to c()

In Splus c() and unlist() called the same C code, but with a different 'sys_index' code (the last argument to .Internal) and c() did not consider an argument named 'use.names' special. > c function(..., recursive = F) .Internal(c(..., recursive = recursive), "S_unlist", TRUE, 1) > unlist

Re: [Rd] error handling in strcapture

Lawrence <lawrence.mich...@gene.com > wrote: > Hi Bill, > > Thanks, another good suggestion. strcapture() now returns NAs for > non-matches. It's nice to have someone kicking the tires on that > function. > > Michael > > On Wed, Sep 21, 2016 at 12:11 PM, Willia

[Rd] error handling in strcapture

Michael, thanks for looking at my first issue with utils::strcapture. Another issue is how it deals with lines that don't match the pattern. Currently it gives an error > strcapture("(.+) (.+)", c("One 1", "noSpaceInLine", "Three 3"), proto=list(Name="", Number=0)) Error in strcapture("(.+)

[Rd] strcapture enhancement

The new strcapture function in R-devel is handy, capturing the matches to the parenthesized subpatterns in a regular expression in the columns of a data.frame, whose column names and classes are given by the 'proto' argument. E.g., > p1 <- data.frame(Name="", Number=0) >

Re: [Rd] Different results for tan(pi/2) and tanpi(1/2)

Other examples of functions like this are log1p(x), which is log(1+x) accurate for small x, and expm1(x), which is exp(x)-1 accurate for small x. E.g., > log1p( 1e-20 ) [1] 1e-20 > log( 1 + 1e-20 ) [1] 0 log itself cannot be accurate here because the problem is that 1 == 1 + 1e-20 in

Re: [Rd] Different results for tan(pi/2) and tanpi(1/2)

tanpi(x) should be more accurate than tan(pi*x), especially near multiples of pi/2. Bill Dunlap TIBCO Software wdunlap tibco.com On Fri, Sep 9, 2016 at 11:55 AM, Hans W Borchers wrote: > The same argument would hold for tan(pi/2). > I don't say the result 'NaN' is wrong,

Re: [Rd] Different results for tan(pi/2) and tanpi(1/2)

It should be the case that tan(pi*x) != tanpi(x) in many cases - that is why it was added. The limits from below and below of the real function tan(pi*x) as x approaches 1/2 are different, +Inf and -Inf, so the limit is not well defined. Hence the computer function tanpi(1/2) ought to return

Re: [Rd] R (development) changes in arith, logic, relop with (0-extent) arrays

Prior to the mid-1990s, S did "length-0 OP length-n -> rep(NA, n)" and it was changed to "length-0 OP length-n -> length-0" to avoid lots of problems like any(x<0) being NA when length(x)==0. Yes, people could code defensively by putting lots of if(length(x)==0)... in their code, but that is

Re: [Rd] R (development) changes in arith, logic, relop with (0-extent) arrays

Shouldn't binary operators (arithmetic and logical) should throw an error when one operand is NULL (or other type that doesn't make sense)? This is a different case than a zero-length operand of a legitimate type. E.g., any(x < 0) should return FALSE if x is number-like and length(x)==0 but

Re: [Rd] withAutoprint({ .... }) ?

Re withAutoprint(), Splus's source() function could take a expression (literal or not) in place of a file name or text so it could support withAutoprint-like functionality in its GUI. E.g., > source(auto.print=TRUE, exprs.literal= { x <- 3:7 ; sum(x) ; y <- log(x) ; x - 100}, prompt="--> ") -->

Re: [Rd] A bug in the R Mersenne Twister (RNG) code?

Try comparing the streams for when the 625-integer versions of the seeds are identical. (R's seed is 626 integers: omit the first value, which indicates which random number generator the seed is for.). I find the the MKL Mersenne Twister results match R's (with occassional differences in the

[Rd] findInterval(all.inside=TRUE) for degenerate 'vec' arguments

What should findInterval(x,vec,all.inside=TRUE) return when length(vec)<=1, so there are no inside intervals? R-3.3.0 gives a decreasing map of x->output when length(vec)==1 and -1's when length(vec)==0. Would '0' in all those cases be better? > findInterval(x=c(10, 11, 12), vec=11,

Re: [Rd] Model object, when generated in a function, saves entire environment when saved

Another solution is to only save the parts of the model object that interest you. As long as they don't include the formula (which is what drags along the environment it was created in), you will save space. E.g., tfun2 <- function(subset) { junk <- 1:1e6 list(subset=subset,

Re: [Rd] Model object, when generated in a function, saves entire environment when saved

One way around this problem is to make a new environment whose parent environment is .GlobalEnv and which contains only what the the call to lm() requires and to compute lm() in that environment. E.g., tfun1 <- function (subset) { junk <- 1:1e+06 env <- new.env(parent = globalenv())

Re: [Rd] Detecting user interrupts in R_tryEval

In R code tryCatch can detect the difference. Hit control-C (on Unixen) or Escape (on Windows) to interrupt the long-running for loop and see that the interrupt clause gets called: > z <- tryCatch(for(i in seq_len(1e8))log(exp(i/10)), error=function(e)e, interrupt=function(e)e) ^C> dput(z)

[Rd] odd warning unlinking symlink on Windows

While constructing some tests of symbolic link code in R, I got an odd warning when trying the remove a symbolic link: file.create(tfile <- tempfile()) #[1] TRUE file.symlink(tfile, tlink <- tempfile()) #[1] TRUE unlink(tlink) #Warning message: #In unlink(tlink) : # cannot delete reparse point

[Rd] meaning of 'Depends: R (> 3.0.2)'

In knitr-1.13 the DESCRIPTION file's Depends line uses a ">" instead of the usual ">=". Depends: R (> 3.0.2) I don't see the strict greater than in Writing R Extensions. Does it really mean that 3.0.2 is not suitable and that whatever version comes after it is? Bill Dunlap TIBCO Software

Re: [Rd] Single-threaded aspect

The R language itself has features that limit how much mulitthreading/parallel processing can be done. There are functions with side effects, such as library(), plot(), runif(), <-, and <<- and there are no mechanisms to isolate them. Bill Dunlap TIBCO Software wdunlap tibco.com On Thu, May 12,

Re: [Rd] "cophenetic" function for objects of class "dendrogram"

Note that cophenetic.default (which works on the output of hclust(dist(X))) uses the row names of X as labels. as.dendrogram.hclust does not retain those row names so cophenetic.dendrogram cannot use them (so it orders them based on the topology of the dendrogram). Bill Dunlap TIBCO Software

Re: [Rd] "cophenetic" function for objects of class "dendrogram"

I think the results differ only in the order of the labels. The following function puts the labels in a standard order and then the results are the same: canonicalize.dist <- function (distObject) { o <- order(labels(distObject)) as.matrix(distObject)[o, o, drop = FALSE] }

[Rd] locked environments

Shouldn't the following 4 ways to alter an object in a locked environment either all work or all fail? (All working would be nice, I think.) E <- new.env() assign("var", c(1,2,3,4), envir=E) lockEnvironment(E, bindings=FALSE) E$var[1] <- 101 ; E$var #[1] 101 2 3

Re: [Rd] (no) circular dependency

> but this strategy quickly inflates the number of packages on CRAN. CRAN contains 8210 packages today, so I would not worry about adding an extra one. Also, I think several small packages are preferable to one large one because attaching a big one just to get the one or two functions you want

Re: [Rd] Optimization bug when byte compiling with gcc 5.3.0 on windows

>If I recall correctly, some eigen vectors had their >direction flipped (negative values became positive and vice versa). >Did you notice anything of this kind when running 'make check' and >'make check recommended' ? It is important to us that numeric results >are reproducible between versions of

[Rd] keepNA in nchar and nzchar

Is it intended that in yesterday's version of R-devel the default value of keepNA is different in nchar (NA) and nzchar (FALSE)? > args(nchar) function (x, type = "chars", allowNA = FALSE, keepNA = NA) NULL > args(nzchar) function (x, keepNA = FALSE) NULL Is it intended that for

[Rd] smooth()'s output attribute 'endrule'?

What is the point of the 'endrule' attribute of smooth()'s return value? For the simpler smooths (not involving 'S'?) it is value of the endrule argument, for more complicated smooths it is the endrule not asked for (the choices are 'Tukey' and 'copy'). For kind="S", it is not in the return

[Rd] as.data.frame and illegal row.names argument (bug in package:DoE.wrapper?)

as.data.frame methods behave inconsistently when they are given a row.name argument of the wrong length. The matrix method silently ignores row.names if it has the wrong length and the numeric, integer, and character methods do not bother to check and thus make an illegal data.frame. >

[Rd] array() ignores illegal non-list dimnames

Is there a reason that array() silently ignores dimnames that are not a list but matrix() gives an error? > str(matrix(11:14, 2, 2, dimnames=c("Rows","Cols"))) Error in matrix(11:14, 2, 2, dimnames = c("Rows", "Cols")) : 'dimnames' must be a list > str(array(11:14, dim=c(2, 2),

<    1   2