根据阈值选择R中的行

Question

In R, I have a matrix with N columns of all numbers. (Each row has a name, but that's irrelevant.) I'd like to return rows where there is at least one column has a value greater than some threshold. Right now, I'm doing something like this:

THRESHOLD <- 10
#  my_matrix[,1] can be ignored
my_matrix <- subset (my_matrix, my_matrix[,1] > THRESHOLD | my_matrix[,2] > THRESHOLD | ... )

It seems odd to have to manually list each column. Also, if the number of input columns changes, I have to rewrite this.

There has to be a better way, but I can't figure out what I should be looking for.

I can convert my matrix to a data frame, if that is easier... Any suggestions would be appreciated!

Answer 1

find any row values greater than threshold using apply and use it to extract the rows from mat data.

mat[apply( mat2, 1, function( x ) any( x > threshold ) ), ]

EDIT:

Break down of the above single line.

# create sample data by simulating samples from standard normal distribution
set.seed(1L)   # set random number generator for consistent data simulation

mat <- matrix( data = c(letters[1:3], as.character( rnorm(9, mean = 0, sd = 1))),
               byrow = FALSE, 
               nrow = 3, 
               ncol = 4 ) # create simulated data matrix

threshold <- 0  # set threshold

mat2 <- apply( mat[, 2:ncol(mat) ], 2, as.numeric )  # extract columns 2 to end and convert to numeric

# Get the logical indices (true or false) if any row has values greater than 0 (threshold)
row_indices <- apply( mat2, 1, function( x ) any( x > threshold ) )

mat[row_indices, ]  # extract matrix data rows that has TRUE in row_indices
#     [,1]                 [,2]                 [,3]                 [,4]               
# [1,] "a"  "-0.626453810742332" "1.59528080213779"   "0.487429052428485"
# [2,] "b"  "0.183643324222082"  "0.329507771815361"  "0.738324705129217"
# [3,] "c"  "-0.835628612410047" "-0.820468384118015" "0.575781351653492"

Note:

In your question, you mentioned that first column is character and the rest are numbers. By rule, matrix can hold one data type. Given this information, I assume that your data matrix is a character data type. You can find it by using class(mat) . If it is character matrix, then extract columns 2 to end and then convert it to numeric. Then use it in the apply loop to check for any values greater than threshold.