InsertRow increases dimensions of scaled vectors

[tkasci]

When applying InsertRow on a data frame containing a variable that is the result of scale(), its dimensions are increased, leading to possibly strange results.
Example:

test <- data.frame(matrix(rnorm(100), ncol=5))
test$X1s <- scale(test$X1)
InsertRow(test, NewRow=(rep(1,6)), RowNum = 15) -> test
dim(test)
#> [1] 21  6
dim(scale(test)) # ???
#> [1] 21 11

[jmcastagnetto]

Changing the implementation from

DataCombine/R/utils.R

Line 272 in 6202157

InsertRow <- function(data, NewRow, RowNum = NULL) {

to a more naive one works on R 3.6:

InsertRow <- function(data, NewRow, RowNum = NULL) {
  if (ncol(data) != length(NewRow)) {
    stop('NewRow must be the same length as the number of columns in the data.\n',
         call. = FALSE)
  }
  if (!is.null(RowNum)) {
    data <- rbind(
      rbind(
        data[1:(RowNum - 1), ],
        NewRow
      ), 
      data[RowNum:nrow(data), ]
      )
  #   data[seq(RowNum + 1,nrow(data) + 1), ] <- data[seq(RowNum, nrow(data)), ]
  #   data[RowNum, ] <- NewRow
  }
  else if (is.null(RowNum)) {
    data <- rbind(data, NewRow)
  }
  return(data)
}

test <- data.frame(matrix(rnorm(100), ncol=5))
test$X1s <- scale(test$X1)
test1 <- InsertRow(test, NewRow=rep(1,6), RowNum = 15)
dim(test)
# [1] 20  6
dim(test1)
# [1] 21  6
dim(scale(test1))
# [1] 21  6

The data looks OK too:

> test[11:16,]
             X1         X2         X3         X4
11 -0.554802046 -0.6853903 -2.0467622  1.4041387
12  0.375084862  0.6801423 -0.6298347  0.2553870
13 -0.862548701 -1.2354197 -0.2718516 -0.7419452
14  1.516296694  2.5151244  2.5458425  1.0323457
15  0.119888979 -0.9082973  0.2822750 -2.1882965
16 -0.007863882 -0.8383689  0.4936497 -1.4547707
            X5        X1s
11  0.29420542 -0.5549986
12 -0.04913802  0.2579263
13  0.81719635 -0.8240366
14 -0.21627068  1.2555956
15  0.97142051  0.0348292
16  0.48755283 -0.0768548
> test1[11:16,]
            X1         X2         X3         X4
11  -0.5548020 -0.6853903 -2.0467622  1.4041387
12   0.3750849  0.6801423 -0.6298347  0.2553870
13  -0.8625487 -1.2354197 -0.2718516 -0.7419452
14   1.5162967  2.5151244  2.5458425  1.0323457
15   1.0000000  1.0000000  1.0000000  1.0000000
151  0.1198890 -0.9082973  0.2822750 -2.1882965
             X5        X1s
11   0.29420542 -0.5549986
12  -0.04913802  0.2579263
13   0.81719635 -0.8240366
14  -0.21627068  1.2555956
15   1.00000000  1.0000000
151  0.97142051  0.0348292
> scale(test1[11:16,])
            X1         X2          X3
11  -0.9079719 -0.6362247 -1.41042885
12   0.1211044  0.3152606 -0.49928659
13  -1.2485454 -1.0194780 -0.26908888
14   1.3840474  1.5938523  1.54280351
15   0.8126782  0.5381333  0.54876365
151 -0.1613128 -0.7915434  0.08723717
             X4         X5         X1s
11   0.93423078 -0.3285733 -0.90890950
12   0.09395609 -0.9718856  0.07640906
13  -0.63556016  0.6513390 -1.23500130
14   0.66227622 -1.2850370  1.28564987
15   0.63861641  0.9938527  0.97585123
151 -1.69351932  0.9403042 -0.19399935
attr(,"scaled:center")
       X1        X2        X3        X4 
0.2656533 0.2276932 0.1466115 0.1269383 
       X5       X1s 
0.4695689 0.1948860 
attr(,"scaled:scale")
       X1        X2        X3        X4 
0.9036131 1.4351589 1.5551112 1.3671145 
       X5       X1s 
0.5337119 0.8250377

Although the row names get a bit funky