熊猫不更新 CSV

Question

数据集： https://github.com/Bene939/newsheadlinedatasets

通过我的程序，我正在标记我的新闻标题数据集。直到今天它工作得很好。 由于某种原因，它不会再写入 csv 文件了。据我所知，数据框已更新。

在我的 csv 的大约 4469 行处，它开始不覆盖 csv 文件。然后它做到了。然后没有再做一次，直到它在第 4474 行完全停止覆盖。直到现在它工作正常，如果我创建一个新的 csv，它将覆盖它。

我正在使用 Jupyter Notebook。这有什么限制吗？带标签的数据集大约 300KB。

!pip install pandas
!pip install pathlib

import pandas as pd
from pathlib import Path

#takes data frame and file name & appends it to given csv
def append_df(df, file_name):
  my_file = Path(file_name)
  if my_file.exists():
    print("Appending to existing file named " + file_name)
    orig_df = pd.read_csv(file_name)
    print("Old Data Frame: ")
    print(orig_df)
    new_df = pd.concat([orig_df, df], ignore_index=True).drop_duplicates()
    print("New Data Frame: ")
    print(new_df)
    new_df.to_csv(file_name, index=False, header = True, encoding='utf-8-sig')
  else:
    print("Creating new file named" + file_name)
    news_sentiment_df.to_csv(file_name, index=False, header = True, encoding='utf-8-sig')

#takes data frame and file name & overwrites given csv 
def update_csv(df, file_name):
  print("Overwriting " + file_name)
  df.to_csv(file_name, index=False, header = True, encoding='utf-8-sig')

#shows sentence by sentence, labels it according to input and saves it in a new csv file
print("WARNING: EDITING CSV FILE WITH EXCEL MAY CORRUPT FILE\n")
file_name = "news_headlines.csv"
new_file = "news_headlines_sentiment.csv"
news_sentiment_df = pd.DataFrame(columns=["news", "sentiment"])
my_file = Path(file_name)
if my_file.exists():
  df = pd.read_csv(file_name, encoding='utf-8-sig', error_bad_lines=False)
  print("Loaded " + file_name)
  for index, row in df.iterrows():
    user_input = -1
    range = [0, 1, 2]
    while user_input not in range:
      print("####################################################################")
      print(row["news"])
      try:
        user_input = int(input("Negative: 0\nNeutral: 1\nPositive: 2\n"))
      except ValueError as err:
        print("\nPlease enter an Integer!\n")
        pass

    new_element = 0
    #label sentiment according to input
    if user_input == 0:
      new_element = [row["news"], 0]
    elif user_input == 1:
      new_element = [row["news"], 1]
    elif user_input == 2:
      new_element = [row["news"], 2]
      

    #save labeled sentence to new file
    news_sentiment_df.loc[len(news_sentiment_df)] = new_element
    append_df(news_sentiment_df, new_file)

    #delete data point from original data frame
    index_name = df[df["news"] == row["news"]].index
    df.drop(index_name, inplace=True)

    #update old csv file
    update_csv(df, file_name)


else:
  print("File not Found")

Answer 1

我在使用 drop_duplicates 函数时尝试添加重复项而没有注意到它