求助：编写R语言函数筛选满足指定非零列数的矩阵行

Hey there! Let's work through this R problem together. I'll show you how to build a custom function that filters your matrix to keep only rows with at least a specified number of columns containing values greater than 0.

Step 1: First, let's recreate your input matrix #

This will let us test the function directly with your data:

# Build the example matrix from your question
mat <- matrix(
  data = c(0, 0, 0, 2,
           2, 5, 4, 0,
           0, 0, 8, 3),
  nrow = 3,
  byrow = TRUE,
  dimnames = list(c("row1", "row2", "row3"),
                  c("Col1", "Col2", "Col3", "Col4"))
)

Step 2: Create the custom filtering function #

This function takes two arguments: your matrix, and the minimum number of positive-value columns you want to require:

filter_rows_by_positive_cols <- function(matrix_input, min_positive_cols) {
  # Calculate how many columns have values > 0 for each row
  positive_col_count <- rowSums(matrix_input > 0)
  
  # Keep only rows where the count meets or exceeds your minimum
  matrix_input[positive_col_count >= min_positive_cols, , drop = FALSE]
}

Step 3: Test the function with your requirements #

Since your desired output keeps rows with at least 2 positive columns, we'll call the function with min_positive_cols = 2:

# Run the filter
filtered_matrix <- filter_rows_by_positive_cols(mat, min_positive_cols = 2)

# View the result
filtered_matrix

What you'll get (matches your expected output): #

Col1 Col2 Col3 Col4
row2    2    5    4    0
row3    0    0    8    3

A quick breakdown of how this works: #

• rowSums(matrix_input > 0): This first creates a boolean matrix where each cell is TRUE if the value is greater than 0, then sums those TRUE values per row to get a count of positive columns for each row.
• The drop = FALSE argument ensures the result stays a matrix (even if only one row is left, it won't convert to a vector accidentally).

Bonus tweaks if you need them: #

• If you ever need to filter for values greater than a different threshold (not 0), just replace matrix_input > 0 with matrix_input > your_threshold (e.g., matrix_input > 5).
• If your matrix has NA values, add na.rm = TRUE to the rowSums call to ignore them: rowSums(matrix_input > 0, na.rm = TRUE).

内容的提问来源于stack exchange，提问作者Erika