R语言For循环创建变量报错及合并数据帧问题求助

Hey there! Let's work through your problem step by step—first fixing that error, then making your code cleaner and more efficient.

First: Why You're Getting That Error #

You hit the nail on the head: nam is just a string (like "cik_0000320193"), not the actual data frame you created with assign(). When you try nam[i,] <- ..., R doesn't recognize it as a data frame—it thinks you're trying to subscript a string, which doesn't make sense.

A quick fix for your existing loop would be to explicitly retrieve the data frame, modify it, then save it back:

for (j in cik_sample){
 c_filings = company_filings(j, type = "10-K",count = 1000 )
 row_num = nrow(c_filings)
 if (row_num == 0){
 next
 }
 c_filings_short = select(c_filings, filing_date, href, type)
 c_hrefs <- as.data.frame(c_filings_short[,2])
 c_length = length(c_hrefs)
 c_index = (1:c_length)
 nam <- paste("cik_", j,sep="")
 assign(nam, data.frame(matrix(ncol =21, nrow = c_length )))
 
 # Fix the inner loop here
 for (i in c_index) {
   # Get the actual data frame from the global environment
   current_df <- get(nam)
   # Fill the ith row
   current_df[i,] <- filing_filers(as.character(c_hrefs[i,1]))
   # Save the updated data frame back
   assign(nam, current_df)
 }
}

But honestly, using assign() and managing separate data frames for each CIK is messy and error-prone. Let's look at a better approach.

Better Approach: Use Lists + Tidyverse Tools #

Instead of creating separate data frames, we can store results in a list and then combine everything into one big data frame automatically. This uses purrr for iterating and dplyr for data manipulation—way cleaner!

First, load the required packages:

library(EDGARWebR)
library(dplyr)
library(purrr)
library(lubridate) # For extracting years from dates

Then write a function to process a single CIK:

process_single_cik <- function(cik) {
  # Get all 10-K filings for the CIK
  filings <- company_filings(cik, type = "10-K", count = 1000)
  
  # Skip if no filings exist
  if (nrow(filings) == 0) {
    return(tibble()) # Return empty tibble to keep things consistent
  }
  
  # Keep only the columns we need, add CIK and year from filing date
  filings_clean <- filings %>%
    select(filing_date, href) %>%
    mutate(
      cik = cik,
      filing_year = year(filing_date)
    )
  
  # Safely fetch filer data (handle cases where a link fails)
  safe_get_filers <- possibly(filing_filers, otherwise = tibble())
  
  # Iterate over each href, get filer data, and combine with filing info
  filings_clean %>%
    mutate(filer_data = map(href, safe_get_filers)) %>%
    unnest(filer_data, keep_empty = TRUE) # Keep rows even if filer data is missing
}

Now process all your CIKs and combine into one data frame:

cik_sample <- c("0001438823","0000320193","0000018230")

# Process every CIK and row-bind the results
final_combined_df <- map_dfr(cik_sample, process_single_cik)

Why This Is Better #

• No more assign() or messy global environment variables—all results are managed in lists/data frames.
• map_dfr() automatically combines all results into a single data frame, so you don't have to manually merge later.
• possibly() adds error handling: if a single filing_filers() call fails (e.g., broken link), it won't crash the entire process.
• We added a filing_year column to make it easier to track which year each filing corresponds to.

Merging Your Existing Data Frames (If You Still Want To) #

If you already ran your original loop and have those cik_xxxx data frames, you can combine them like this:

# Get all data frame names starting with "cik_"
cik_df_names <- ls(pattern = "^cik_")

# Combine them into one data frame
combined_df <- bind_rows(mget(cik_df_names))

内容的提问来源于stack exchange，提问作者jayjunior