similar to: Sum of columns of a data frame equal to NA when all the elements are NA

On 21/03/2018 11:44 AM, Stefano Sofia wrote: > Dear list users, > let me ask you this trivial question. I worked on that for a long time, by now. > Suppose to have a data frame with NAs and to sum some columns with rowSums: > > df <- data.frame(A = runif(10), B = runif(10), C = rnorm(10)) > df[1, ] <- NA > rowSums(df[ , which(names(df) %in%

What do you mean by "should not"? NULL means "missing object" in R. The result of the sum function is always expected to be numeric... so NA_real or NA_integer could make sense as possible return values. But you cannot compute on NULL so no, that doesn't work. See the note under the "Value" section of ?sum as to why zero is returned when all inputs are removed.

No. The empty sum is zero. Adding it to another sum should not change it. Nothing audacious about that. This is consistent; other definitions just cause trouble. -pd > On 21 Mar 2018, at 18:05 , Boris Steipe <boris.steipe at utoronto.ca> wrote: > > Surely the result of summation of non-existent values is not defined, is it not? And since the NA values have been _removed_,

Should not the result be NULL if you have removed the NA with na.rm=TRUE ? B. > On Mar 21, 2018, at 11:44 AM, Stefano Sofia <stefano.sofia at regione.marche.it> wrote: > > Dear list users, > let me ask you this trivial question. I worked on that for a long time, by now. > Suppose to have a data frame with NAs and to sum some columns with rowSums: > > df <-

Surely the result of summation of non-existent values is not defined, is it not? And since the NA values have been _removed_, there's nothing left to sum over. In fact, pretending the the result in that case is zero would appear audacious, no? Cheers, Boris > On Mar 21, 2018, at 12:58 PM, Jeff Newmiller <jdnewmil at dcn.davis.ca.us> wrote: > > What do you mean by

I see: consistency with additive identity. That makes sense. Thanks. B. > On Mar 21, 2018, at 1:22 PM, peter dalgaard <pdalgd at gmail.com> wrote: > > No. The empty sum is zero. Adding it to another sum should not change it. Nothing audacious about that. This is consistent; other definitions just cause trouble. > > -pd > >> On 21 Mar 2018, at 18:05 , Boris

"I see: consistency with additive identity. " Ummm, well: > 1+NULL numeric(0) > sum(1,NULL) [1] 1 Of course, there could well be something here I don't get, but that doesn't look very consistent to me. However, as I said privately, so long as the corner case behavior is documented, which it is, I don't care. Cheers, Bert Bert Gunter "The trouble with

Dear list users, I have to aggregate wind direction data (wd) using a function that requires also a second input variable, wind speed (ws). This is the function that I need to use: my_fun <- function(wd1, ws1){ u_component <- -ws1*sin(2*pi*wd1/360) v_component <- -ws1*cos(2*pi*wd1/360) mean_u <- mean(u_component, na.rm=T) mean_v <- mean(v_component, na.rm=T) mean_wd

Dear R-users, suppose I have three dataframes like these df1: mydate min_temp 31032013 12 01042013 8 02042013 -999 df2: mydate min_temp 31032013 10 01042013 11 02042013 14 df3: mydate min_temp 31032013 4 01042013 3 02042013 5 where -999 means that the temperature data is not available (at the moment I cannot change it to NA because I am not the db administrator); suppose also that oggi is

?s 15:51 de 13/05/2023, Stefano Sofia escreveu: > Dear list users, > > I have to aggregate wind direction data (wd) using a function that requires also a second input variable, wind speed (ws). > > This is the function that I need to use: > > > my_fun <- function(wd1, ws1){ > > u_component <- -ws1*sin(2*pi*wd1/360) > v_component <-

Dear R-list users, I need to use strptime because I have to deal with date with hours and minutes. I read the manual for strptime and I also looked at many examples, but when I try to apply it to my code, I always encounter some problems. I try to change the default format, with no success. Why? How can I change the format? 1. init_day <- as.factor("2015-02-24-00-30")

Dear Jim and dear Enrico, thank you for your replies. Unfortunately your hints didn't solve my problem, and I am getting mad. Can I show you my whole process? I will be as quick as possible. I start from a data frame called Snow of the form year month day hh mm hs 2007 11 19 0 0 0.00 2007 11 19 0 30 0.00 2007 11 19 1 0 0.00 2007 11 19 1 30 0.00 2007 11 19 2 0 0.00 2007 11 19 2 30 0.00 2007 11

Hi Stefano, As the help page says: "The default for the format methods is "%Y-%m-%d %H:%M:%S" if any element has a time component which is not midnight, and "%Y-%m-%d" otherwise. This is because when the result is printed, it uses the default format. If you want a specified output representation: format(strptime(init_day, format="%Y-%m-%d-%H-%M"),"%Y-%M-%d

Dear list users, if within a function I first define a matrix mymat <- matrix(NA, nrow=m, ncol=n) and somewhere afterwards by mistake mymat <- c(1,2,3) this second command deletes the first one. How can I make sure that within the function mymat will always remain the matrix defined at the beginning, without possibility of changing it throughout the code? Secondly, is it possible to define

Dear R list users, the singluar value decomposition of a symmetric matrix M is UDV^(T), where U = V. La.svd(M) gives as output three elements: the diagonal of D and the two orthogonal matrices u and vt (which is already the transpose of v). I noticed that the transpose of vt is not exactly u. Why is that? thank you for your attention and your help Stefano AVVISO IMPORTANTE: Questo messaggio di

Fwiw the engine behind geom_raster needs explicit observation-per-row form for input (with no structural normalization), so conversion to points is perfectly proper here, albeit confusing in context. (It's closer to what graphics devices actually use ultimately, but the expansion is laid out very early in ggplot2 because there's no standard for intermediate forms.) Cheers, Mike On Wed,

I would like to simulate some SARIMA models, e.g. a SARIMA (1,0,1)(1,0,1)[4] process. I installed the package 'forecast', where the function simulate.Arima should do what I am trying to do. I am not able to understand how it works Could somebody help me with an example? thank you Stefano Sofia AVVISO IMPORTANTE: Questo messaggio di posta elettronica pu? contenere informazioni

1. Martin Maechler's comments should be taken as replacements for anything I wrote where appropriate. Any apparent conflict is a result of his superior knowledge. 2. 'eigen' returns the eigenvalue decomposition assuming the matrix is symmetric, ignoring anything in m[upper.tri(m)]. 3. The basic idea behind both posdefify and nearPD is to compute the

Dear R users, I have difficulty to create a list within a list. Example: with > A <- vector(mode="list", 4) I create a list of 4 elements: > A [[1]] NULL [[2]] NULL [[3]] NULL [[4]] NULL In each element of this list I can store, for example, a matrix: A[[1]] <- matrix ... I need each element of A to be a list (all the lists of the same length), and then store matrices

Hello, I would like use Kalman filter for estimating parameters of a stochastic model. I have developed the state space model but I don’t know the correct way use Kalman filter for parameter estimation. Has anybody experience in work with Kalman filter in R. I don’t know the correct function. Maybe it is - KalmanLike; but what is the correct Input? - tsmooth? -