如何按行计算标准差？rowSds函数执行失败问题求助

Hey there! Let's troubleshoot why your rowSds() call isn't working even though rowMeans() runs fine. Here are the most likely issues and fixes:

1. You're Missing the matrixStats Package #

First off, rowSds() isn't part of R's base packages—it belongs to the matrixStats library. If you haven't installed or loaded this package yet, R will throw a "could not find function 'rowSds'" error.

Fix this by:

# Install the package (only need to do this once)
install.packages("matrixStats")

# Load it before using rowSds()
library(matrixStats)

2. Your Data Might Be a Data Frame Instead of a Matrix #

While rowMeans() works seamlessly with data frames, rowSds() is optimized for matrix inputs. If your values_for_all object is a data frame, converting it to a matrix will resolve most issues.

Modify your code like this:

# Convert data frame to matrix
values_matrix <- as.matrix(values_for_all)

# Now calculate row-wise SD
SD <- rowSds(values_matrix, na.rm = TRUE)

3. Alternative: Use Base R (No Extra Package Needed) #

If you'd rather avoid installing additional packages, you can use apply() with base R's sd() function to get the same result:

SD <- apply(values_for_all, 1, function(x) sd(x, na.rm = TRUE))

This applies the sd() function to each row (the 1 in the second argument specifies rows) and ignores NA values.

Test with Your Sample Data #

Let's verify with your input example to make sure it works:

# Recreate your sample data
values_for_all <- data.frame(
  V1 = c(1, 2),
  V2 = c(0.35, 8.75),
  V3 = c(0.35, 3.22),
  V4 = c(4.33, 7.34),
  V5 = c(0.09, 0.56),
  V6 = c(4.17, 2.43),
  V7 = c(0.16, 0.23),
  V8 = c(9.90, 1.20),
  V9 = c(15.25, 8.45),
  V10 = c(0.16, 1.26),
  V11 = c(2.38, NA),
  V12 = c(2.55, NA),
  V13 = c(8.14, 1.24),
  V14 = c(0.16, 0.16),
  V15 = c(NA, 2.34),
  V16 = c(0.16, 0.36)
)

# Clean and convert data (your original code)
values_for_all[values_for_all == ''] <- NA
values_for_all[] <- lapply(values_for_all, as.numeric)

# Calculate means (works as expected)
values_mean <- rowMeans(values_for_all, na.rm=TRUE)

# Calculate SD with matrixStats
library(matrixStats)
SD_matrixStats <- rowSds(as.matrix(values_for_all), na.rm=TRUE)

# Calculate SD with base R
SD_baseR <- apply(values_for_all, 1, function(x) sd(x, na.rm = TRUE))

# Both methods give the same result
print(SD_matrixStats)
print(SD_baseR)

Running this should output the correct row-wise standard deviations for your dataset.

内容的提问来源于stack exchange，提问作者student24