基于Scopus通过DOI获取文献参考文献的R代码执行问题

Hey there, I see you're having trouble pulling reference titles using the rscopus package—let's get this sorted out without messy eval(parse) hacks, which are error-prone and generally not the best practice here.

The Core Issue with Your Original Code #

Your approach of building a string to evaluate was overcomplicating things. The abstract_retrieval response returns a nested list structure, so we can directly index into it instead of constructing code as text. Let's simplify this.

Working Solution Code #

Here's a cleaned-up, reliable version of your code that properly extracts reference titles and stores them in your data frame:

library(rscopus)

# Load your sample DOI data
df <- data.frame(doi = c("10.1109/TPAMI.2018.2798607", "10.1109/CNS.2017.8228696"))

# Initialize empty results data frame
df_references <- data.frame(
  initial_paper = character(),
  ref_title = character(),
  stringsAsFactors = FALSE
)

# Loop through each DOI
for (i in 1:nrow(df)) {
  # Retrieve the abstract/metadata for the current DOI
  retrieval_result <- abstract_retrieval(df$doi[i], identifier = "doi")
  
  # Extract the list of references from the response
  references <- retrieval_result$content$`abstracts-retrieval-response`$`item`$bibrecord$tail$`bibliography`$reference
  
  # Skip if there are no references for this paper
  if (is.null(references)) next
  
  # Loop through each reference to extract the title
  for (ref in references) {
    # Get the title text (handle cases where title might be missing)
    ref_title <- if (!is.null(ref$`ref-info`$`ref-title`$`ref-titletext`)) {
      ref$`ref-info`$`ref-title`$`ref-titletext`
    } else {
      "No title available"
    }
    
    # Add to results data frame
    df_references <- rbind(df_references, data.frame(
      initial_paper = df$doi[i],
      ref_title = ref_title,
      stringsAsFactors = FALSE
    ))
  }
}

# View the results
print(df_references)

Key Improvements Explained: #

• Direct List Indexing: Instead of building strings to evaluate, we directly access the nested list elements (ref$ref-info$ref-title$ref-titletext`). This is cleaner and avoids parsing errors.
• Null Handling: Added checks to skip papers with no references, and handle cases where a reference might not have a title (so your code doesn't break unexpectedly).
• Cleaner Initialization: Properly initialized the results data frame with correct column types (stringsAsFactors = FALSE to avoid outdated factor behavior).

Why eval(parse) Was a Problem #

Constructing code as text and evaluating it is risky because:

• It's easy to make syntax errors in the string (like missing brackets or quotes).
• It can introduce security issues if you're working with untrusted data.
• It makes the code harder to read and debug.

Direct list access is the standard, safer way to work with nested list structures in R.

内容的提问来源于stack exchange，提问作者Elr Mant