在R中按矩阵单元格值删除列并解决下标越界错误

Hey there! Let's work through your R problems step by step—these are super common issues when manipulating matrix data, so you’re in good company.

1. Better Way to Convert the Second Row to Numeric #

Your current approach works, but we can simplify it a lot since data is already a matrix. You don’t need to wrap fly_data[2,] in as.matrix() again—just convert the row directly to numeric:

# Clean up the numeric conversion for row 2
data[2, ] <- as.numeric(data[2, ])

If there’s any chance some values in row 2 can’t be converted to numbers (which would create NAs), you could add a check to handle those later, but this is the most concise and efficient way for your use case.

2. Fixing the "Subscript Out of Bounds" Error in Your For Loop #

The error happens because you’re deleting columns while looping forward through the column indices. Every time you remove a column, the total number of columns in data shrinks, but your loop still uses the original column count. By the time i reaches the original number of columns, that column no longer exists!

There are two solid fixes here:

Option 1: Loop Backwards

Instead of starting at column 1, start at the last column and work your way to the first. This way, deleting a column doesn’t affect the indices of the columns you haven’t checked yet:

# Loop from last column to first to avoid index issues
for (i in ncol(data):1) {
  # Combine your conditions with %in% for cleaner code
  if (data[1, i] %in% c("A", "B") || data[2, i] > 30) {
    data <- data[, -i]
  }
}

Option 2: Vectorized Column Selection (Recommended!)

R shines with vectorized operations—this method is way faster (critical for your 100k+ row matrix) and avoids loops entirely. We’ll first create a logical vector that marks which columns to keep, then subset the matrix in one go:

# Create a logical vector for columns we want to keep
keep_cols <- !(data[1, ] %in% c("A", "B")) & (data[2, ] <= 30)

# Keep only the valid columns
data <- data[, keep_cols]

Pro tip: If row 2 has any values that turned into NA during conversion, you might want to add a check to exclude those (e.g., & !is.na(data[2, ]) to keep_cols) to avoid unexpected behavior.

That should resolve both your issues cleanly!

内容的提问来源于stack exchange，提问作者rbeginner