CRAN Package Check Results for Package orderedLasso

Last updated on 2020-01-20 01:50:28 CET.

Flavor Version Tinstall Tcheck Ttotal Status Flags
r-devel-linux-x86_64-debian-clang 1.7.1 9.75 79.32 89.07 ERROR
r-devel-linux-x86_64-debian-gcc 1.7.1 8.90 64.94 73.84 OK
r-devel-linux-x86_64-fedora-clang 1.7.1 114.40 OK
r-devel-linux-x86_64-fedora-gcc 1.7.1 110.43 OK
r-devel-windows-ix86+x86_64 1.7.1 29.00 97.00 126.00 OK
r-devel-windows-ix86+x86_64-gcc8 1.7.1 18.00 148.00 166.00 OK
r-patched-linux-x86_64 1.7.1 8.69 73.79 82.48 OK
r-patched-solaris-x86 1.7.1 150.80 OK
r-release-linux-x86_64 1.7.1 8.14 74.44 82.58 OK
r-release-windows-ix86+x86_64 1.7.1 28.00 82.00 110.00 OK
r-release-osx-x86_64 1.7.1 OK
r-oldrel-windows-ix86+x86_64 1.7.1 13.00 80.00 93.00 OK
r-oldrel-osx-x86_64 1.7.1 OK

Check Details

Version: 1.7.1
Check: examples
Result: ERROR
    Running examples in 'orderedLasso-Ex.R' failed
    The error most likely occurred in:
    
    > base::assign(".ptime", proc.time(), pos = "CheckExEnv")
    > ### Name: orderedLasso.cv
    > ### Title: Cross-validation function for the ordered lasso
    > ### Aliases: orderedLasso.cv
    >
    > ### ** Examples
    >
    > set.seed(3)
    > n = 50
    > b = c(4,3,1,0)
    > p = length(b)
    > x = matrix(rnorm(n*p),nrow = n)
    > sigma = 5
    > y = x %*% b + sigma * rnorm(n, 0, 1)
    > cvmodel = orderedLasso.cv(x,y, intercept = FALSE, trace = TRUE,
    + method = "Solve.QP", strongly.ordered = TRUE)
    Fold 1 :
    lambda= 0.0555431880114968
    lambda= 0.0648632585178277
    lambda= 0.0757472240282587
    lambda= 0.0884575039721481
    lambda= 0.103300551397943
    lambda= 0.120634241753858
    lambda= 0.14087650149579
    lambda= 0.164515384563759
    lambda= 0.192120839677229
    lambda= 0.22435845216639
    lambda= 0.26200549166382
    lambda= 0.305969652576716
    lambda= 0.357310939184576
    lambda= 0.417267223026156
    lambda= 0.487284088780777
    lambda= 0.569049688247448
    lambda= 0.664535442773709
    lambda= 0.776043575496029
    lambda= 0.906262619424709
    lambda= 1.05833223970905
    lambda= 1.23591893298942
    lambda= 1.44330442899635
    lambda= 1.6854889258164
    lambda= 1.96831164789345
    lambda= 2.29859163349676
    lambda= 2.68429214613239
    lambda= 3.1347126739632
    lambda= 3.6607131464672
    lambda= 4.27497577434272
    lambda= 4.99231082578924
    lambda= 5.83001371162726
    lambda= 6.80828198880995
    lambda= 7.9507023365501
    lambda= 9.28481924637094
    lambda= 10.8427991375648
    lambda= 12.662205910312
    lambda= 14.7869066355452
    lambda= 17.2681292183268
    lambda= 20.165697535686
    lambda= 23.5494738288839
    lambda= 27.5010431271164
    lambda= 32.1156803151964
    lambda= 37.5046472724847
    lambda= 43.7978754685737
    lambda= 51.1470986948335
    lambda= 59.7295114640004
    lambda= 69.752041284182
    lambda= 81.4563378145596
    lambda= 95.124599194552
    lambda= 111.086376022994
     ----------- FAILURE REPORT --------------
     --- failure: the condition has length > 1 ---
     --- srcref ---
    :
     --- package (from environment) ---
    orderedLasso
     --- call from context ---
    predict.orderedLasso(object = object, newdata = newdata, ...)
     --- call from argument ---
    if (class(object$beta) == "numeric") {
     if (is.null(object$b0)) {
     b0 = 0
     if (ordered)
     b0.ordered = 0
     }
     else {
     b0 = object$b0
     if (ordered)
     b0.ordered = object$b0.ordered
     }
     yhat = as.vector(newdata %*% object$beta + b0)
     if (ordered)
     yhat.ordered = as.vector(newdata %*% object$beta.ordered +
     b0.ordered)
    } else {
     if (is.null(object$b0)) {
     b0 = rep(0, length = ncol(object$beta))
     if (ordered)
     b0.ordered = rep(0, length = ncol(object$beta))
     }
     else {
     b0 = object$b0
     if (ordered)
     b0.ordered = object$b0.ordered
     }
     nlam = ncol(object$beta)
     yhat = matrix(NA, n, nlam)
     if (ordered)
     yhat.ordered = matrix(NA, n, nlam)
     for (i in seq(nlam)) {
     yhat[, i] = as.vector(newdata %*% object$beta[, i] +
     b0[i])
     if (ordered)
     yhat.ordered[, i] = as.vector(newdata %*% object$beta.ordered[,
     i] + b0.ordered[i])
     }
    }
     --- R stacktrace ---
    where 1: predict.orderedLasso(object = object, newdata = newdata, ...)
    where 2: predict.orderedLasso.path(a, newdata = x[folds[[ii]], ])
    where 3: orderedLasso.cv(x, y, intercept = FALSE, trace = TRUE, method = "Solve.QP",
     strongly.ordered = TRUE)
    
     --- value of length: 2 type: logical ---
    [1] FALSE FALSE
     --- function from context ---
    function (object, newdata, ...)
    {
     n <- nrow(newdata)
     ordered = object$strongly.ordered
     if (class(object$beta) == "numeric") {
     if (is.null(object$b0)) {
     b0 = 0
     if (ordered)
     b0.ordered = 0
     }
     else {
     b0 = object$b0
     if (ordered)
     b0.ordered = object$b0.ordered
     }
     yhat = as.vector(newdata %*% object$beta + b0)
     if (ordered)
     yhat.ordered = as.vector(newdata %*% object$beta.ordered +
     b0.ordered)
     }
     else {
     if (is.null(object$b0)) {
     b0 = rep(0, length = ncol(object$beta))
     if (ordered)
     b0.ordered = rep(0, length = ncol(object$beta))
     }
     else {
     b0 = object$b0
     if (ordered)
     b0.ordered = object$b0.ordered
     }
     nlam = ncol(object$beta)
     yhat = matrix(NA, n, nlam)
     if (ordered)
     yhat.ordered = matrix(NA, n, nlam)
     for (i in seq(nlam)) {
     yhat[, i] = as.vector(newdata %*% object$beta[, i] +
     b0[i])
     if (ordered)
     yhat.ordered[, i] = as.vector(newdata %*% object$beta.ordered[,
     i] + b0.ordered[i])
     }
     }
     if (!ordered)
     yhat.ordered = NULL
     return(list(yhat = yhat, yhat.ordered = yhat.ordered))
    }
    <bytecode: 0x6c45d58>
    <environment: namespace:orderedLasso>
     --- function search by body ---
    Function predict.orderedLasso in namespace orderedLasso has this body.
     ----------- END OF FAILURE REPORT --------------
    Error in if (class(object$beta) == "numeric") { :
     the condition has length > 1
    Calls: orderedLasso.cv -> predict.orderedLasso.path -> predict.orderedLasso
    Execution halted
Flavor: r-devel-linux-x86_64-debian-clang