如何使用apply函数重写R语言中的for循环

Hey there! Let's work through how to replace your for loop with apply() functions, using your existing data setup as context. First, let's recap your data generation code to make sure we're aligned:

# 你的数据生成函数与数据构建代码
rand.sum <- function(n){ 
  x <- sort(runif(n-1)) 
  c(x,1) - c(0,x) 
} 
M <- t(replicate(8,rand.sum(8))) # 含概率的数据框
Test1 <- as.data.frame(t(replicate(1, rand.sum(8)))) 
Test2 <- as.data.frame(t(replicate(1, rand.sum(8)))) 
Test <- rbind(Test1,Test2) # 测试数据
ID1 <- c("ID1") 
ID2 <- c("ID2") 
ID <- rbind(ID1,ID2) # ID信息
data <- cbind(ID,Test)

第一步：明确你的原始for循环目标 #

Since you didn't share the exact for loop you want to replace, I'll use a common example that fits your data structure: let's say your original loop iterates over each row of data (each ID's test values) to calculate a statistic like row sum, mean, or a custom metric. For example, a typical for loop might look like this:

# 示例原始for循环：计算每行测试数据的总和
row_results <- vector("numeric", nrow(data))
for(i in 1:nrow(data)){
  # 排除第一列的ID，计算该行数值的总和
  row_results[i] <- sum(data[i, -1])
}
data$row_total <- row_results

第二步：用apply()重写循环 #

The apply() function is designed exactly for this kind of row/column-wise operation. Here's how to replace the above loop:

# 用apply计算每行总和（MARGIN=1表示按行应用函数）
data$row_total <- apply(data[, -1], MARGIN = 1, FUN = sum)

扩展：自定义函数的场景 #

If you're using a custom function in your loop, apply() handles that just as easily. For example, if you want to calculate the coefficient of variation (CV = sd/mean) for each row:

# 定义自定义函数
calc_cv <- function(x) {
  sd(x) / mean(x)
}

# 用apply应用到每行
data$row_cv <- apply(data[, -1], MARGIN = 1, FUN = calc_cv)

关键说明 #

• MARGIN=1 tells apply() to operate on rows; use MARGIN=2 if you want to operate on columns instead.
• data[, -1] selects all columns except the first ID column, since we only want to run calculations on the numeric test data.
• apply() is more concise than a for loop, easier to read, and often performs well for moderate-sized datasets.

内容的提问来源于stack exchange，提问作者Mucteam