mwe_lucas.R

# set parameters for the dataset
possible_values <- c(1:10, 99, NA)
n_rows <- 1000
n_cols <- 10

col_names <- letters[1:n_cols]

# create a matrix with the sampled data, that will be transformed into a data.frame in the next step
mwe_matrix <- matrix(sample(possible_values, n_rows * n_cols, replace = T),
                     nrow = n_rows, ncol = n_cols)

# transform into a dataset
df <- as.data.frame(mwe_matrix)
names(df) <- col_names

head(df)
#    a  b  c  d  e  f  g  h  i  j
# 1  6  7  7  5  7 99  5  3 99  3
# 2  4 99  8 99 10  7 99  4 NA  8
# 3 NA  7  9 99  2  4  4  1  5  5
# 4  8  9 10  6  3  2  5 10  5  2
# 5  4  9 99 10  2  4 NA  5  7 99
# 6  1  2  7  3  5  5  2  6  5  3

# Mean per Column
# wrong try 1: with NAs
apply(df, 2, mean)
#  a  b  c  d  e  f  g  h  i  j 
# NA NA NA NA NA NA NA NA NA NA 

# wrong try 2: without NAs but with 99s
apply(df, 2, mean, na.rm = T)
#        a        b        c        d        e        f        g        h        i        j 
# 12.77802 13.66377 14.87859 14.09660 13.15812 12.59477 13.21288 13.84785 14.21098 13.08490 

# compose a new function that excludes certain numbers, it defaults to excluding 99, but you can add more values if you need
mean_fun <- function(x, exclude_nums = c(99)) {
  mean(x[!x %in% exclude_nums], na.rm = T)
}

apply(df, 2, mean_fun)
#        a        b        c        d        e        f        g        h        i        j 
# 5.338061 5.444709 5.600490 5.472793 5.511876 5.462264 5.562426 5.497579 5.671801 5.452581 

# also exclude 10, 9, and 8
apply(df, 2, mean_fun, exclude_nums = c(99, 10, 9, 8))
#        a        b        c        d        e        f        g        h        i        j 
# 3.968801 4.088235 3.851852 3.891798 3.957265 4.023333 4.061329 3.984321 4.015817 3.907216