检查数据结构并填补缺失值：股票数据处理函数开发需求

Hey there! Let's work through this missing value handling task for your stock data data.frame. I'll put together a practical R function that follows your requirements, with clear explanations along the way.

First, let's lock in the core logic we need to implement:

• For each column, identify if it’s continuous (numeric/integer) or categorical (factor/character)
• Calculate the missing value percentage for the column
• Apply your rules:
  • Continuous data: Drop the column if missing rate >40%, else fill missing values with the column mean
  • Categorical data: Drop the column if missing rate >40% (you didn’t specify an alternative, so I’ll use drop as default—super easy to adjust if you want to use mode instead)

Step 1: Full Code Implementation #

Here's the function that handles all this, with helpful print statements to track exactly what's happening to each column:

handle_missing_values <- function(df) {
  # Helper function to calculate mode (for categorical data imputation)
  get_mode <- function(x) {
    unique_vals <- unique(x[!is.na(x)])
    unique_vals[which.max(tabulate(match(x[!is.na(x)], unique_vals)))]
  }

  # Loop through every column in the data frame
  for (col_name in colnames(df)) {
    col_data <- df[[col_name]]
    missing_pct <- sum(is.na(col_data)) / nrow(df)

    # Handle continuous columns (numeric/integer types)
    if (is.numeric(col_data) || is.integer(col_data)) {
      if (missing_pct > 0.4) {
        # Drop column if missing rate exceeds 40%
        df[[col_name]] <- NULL
        cat("✅ Removed continuous column:", col_name, "| Missing rate:", round(missing_pct*100, 2), "%\n")
      } else {
        # Fill missing values with the column's mean
        fill_mean <- mean(col_data, na.rm = TRUE)
        df[[col_name]][is.na(col_data)] <- fill_mean
        cat("🔧 Imputed continuous column:", col_name, "with mean:", round(fill_mean, 2), "| Missing rate:", round(missing_pct*100, 2), "%\n")
      }
    }
    # Handle categorical columns (factor/character types)
    else if (is.factor(col_data) || is.character(col_data)) {
      if (missing_pct > 0.4) {
        # Drop column if missing rate exceeds 40% (swap this with mode code if you prefer to keep it)
        df[[col_name]] <- NULL
        cat("✅ Removed categorical column:", col_name, "| Missing rate:", round(missing_pct*100, 2), "%\n")
      } else {
        # Fill missing values with the column's mode
        fill_mode <- get_mode(col_data)
        df[[col_name]][is.na(col_data)] <- fill_mode
        cat("🔧 Imputed categorical column:", col_name, "with mode:", fill_mode, "| Missing rate:", round(missing_pct*100, 2), "%\n")
      }
    }
    # Optional: Add date column handling (uncomment below if needed)
    # else if (inherits(col_data, "Date") || inherits(col_data, "POSIXct")) {
    #   fill_date <- median(col_data, na.rm = TRUE)
    #   df[[col_name]][is.na(col_data)] <- fill_date
    #   cat("🔧 Imputed date column:", col_name, "with median date:", fill_date, "| Missing rate:", round(missing_pct*100, 2), "%\n")
    # }
  }

  return(df)
}

Step 2: How to Use the Function #

Just pass your stock data frame into the function. For example, if your raw data is stored in stock_df:

# Clean the data using the function
cleaned_stock_data <- handle_missing_values(stock_df)

# Check the first few rows of your cleaned data
head(cleaned_stock_data)

Quick Adjustments You Might Want #

• If you don’t want to drop categorical columns with >40% missing values, replace the df[[col_name]] <- NULL line with the mode imputation code from the else block
• If your data includes date columns, uncomment the date handling section and adjust the fill logic (median date, linear interpolation, etc.) to fit your needs
• Remove the cat() statements if you don’t want verbose output during processing

内容的提问来源于stack exchange，提问作者Sugand Anand