如何在R语言中将数据集转换为矩阵？含示例需求

Hey there! Converting row-based datasets to matrices in R is super straightforward—let me walk you through it with a concrete example, since you mentioned you have a specific dataset and target matrix style in mind.

Let's Start with a Sample Dataset #

First, let's simulate a common row-based dataset you might be working with: each row represents a sample, and columns represent different features (plus an optional sample ID column):

# Simulate a row-oriented dataset
row_data <- data.frame(
  SampleID = c("Sample_A", "Sample_B", "Sample_C"),
  Temp = c(22.5, 24.1, 21.8),
  Humidity = c(65, 72, 58),
  Pressure = c(1013, 1009, 1015)
)

Common Conversion Scenarios #

1. Basic Conversion (Keep Rows as Matrix Rows) #

If your target matrix uses each dataset row as a matrix row (with features as columns), follow these steps:

• First, exclude any non-numeric columns (like SampleID) if you don't want them in the matrix.
• Use as.matrix() to convert the numeric subset to a matrix.
• Optional: Assign row names to the matrix using the sample IDs for clarity.

# Extract numeric columns and convert to matrix
target_matrix <- as.matrix(row_data[, -1])

# Add row names from SampleID
rownames(target_matrix) <- row_data$SampleID

# Check the result
print(target_matrix)

This will output your desired matrix:

Temp Humidity Pressure
Sample_A 22.5       65     1013
Sample_B 24.1       72     1009
Sample_C 21.8       58     1015

2. Transpose the Matrix (Swap Rows and Columns) #

If your target matrix requires rows and columns to be swapped (e.g., features as rows, samples as columns), use the t() function after converting:

transposed_matrix <- t(as.matrix(row_data[, -1]))
rownames(transposed_matrix) <- colnames(row_data[, -1])
colnames(transposed_matrix) <- row_data$SampleID

print(transposed_matrix)

Output:

Sample_A Sample_B Sample_C
Temp          22.5     24.1     21.8
Humidity      65.0     72.0     58.0
Pressure    1013.0   1009.0   1015.0

3. Using dplyr for Column Selection (Optional) #

If you prefer using the dplyr package for cleaner column selection (e.g., selecting all numeric columns automatically), you can do this:

library(dplyr)

target_matrix <- row_data %>%
  select(where(is.numeric)) %>% # Select only numeric columns
  as.matrix()

rownames(target_matrix) <- row_data$SampleID

Customizing for Your Exact Dataset #

If your dataset has a different structure (e.g., nested data, extra metadata columns), you just need to adjust which columns you feed into as.matrix(). The core logic stays the same: isolate the values you want in the matrix, then convert with as.matrix().

内容的提问来源于stack exchange，提问作者Alison