PySimpleGUI调用内部函数后CSV表格列名不更新问题求助

问题背景 #

执行流程：

• 上传CSV文件并创建可编辑的CSV窗口
• 点击「Calculate Diplotype」按钮，调用内部函数diplotype_calc()
• 尝试用处理后的DataFrame更新CSV窗口

问题：数据已按预期处理完成，但CSV窗口始终保留初始上传时的列名，新增的「Metabolizer Status」和「Activity Score」列名无法显示。

相关代码 #

import PySimpleGUI as sg
import webview
import pandas as pd
import csv
import os
import operator
import tkinter as tk
from tkinter import filedialog
import tkinter.messagebox

# Diplotype Calculator Functions
def diplotype_calc(file_path):
    data, allele_cols = read_csv_file(file_path)
    allele_df = pd.DataFrame(data, columns=allele_cols)[['sample ID', 'CYP2D6', 'CYP2C9']]
    print(allele_df)
    # Import CPIC dataframe from Google Cloud Storage into pandas dataframe
    # Currently only supported for CYP2D6 and CYP2C9
    cpic_data = f'https://storage.googleapis.com/pg-genotyping-cpic-2d6-2d9/2d6_2d9_joined.csv'
    cpic_df = pd.read_csv(cpic_data)

    # Merge the two DataFrames based on the common columns CYP2D6 and CYP2C9
    merged_df = pd.merge(allele_df, cpic_df, on=['CYP2D6', 'CYP2C9'])

    # Create two new columns to store the Metabolizer Status and Activity Score
    merged_df['Metabolizer Status'] = ''
    merged_df['Activity Score'] = ''

    # Create two dictionaries to store the metabolizer status and activity score for each call
    d6_status_dict = dict(zip(cpic_df['CYP2D6'], cpic_df['Metabolizer Status']))
    d9_status_dict = dict(zip(cpic_df['CYP2C9'], cpic_df['Metabolizer Status']))
    d6_activity_dict = dict(zip(cpic_df['CYP2D6'], cpic_df['Activity Score']))
    d9_activity_dict = dict(zip(cpic_df['CYP2C9'], cpic_df['Activity Score']))

    # Create two new columns to store the Metabolizer Status and Activity Score
    allele_df['Metabolizer Status'] = ''
    allele_df['Activity Score'] = ''

    # Iterate over each row of the allele DataFrame
    for index, row in allele_df.iterrows():
        # Get the cyp2d6 and cyp2c9 calls for the current sample ID
        cyp2d6_call = row['CYP2D6']
        cyp2c9_call = row['CYP2C9']

        # Use the dictionaries to get the metabolizer status and activity score for the current calls
        cyp2d6_status = d6_status_dict.get(cyp2d6_call, 'UND')
        cyp2c9_status = d9_status_dict.get(cyp2c9_call, 'UND')
        cyp2d6_activity = d6_activity_dict.get(cyp2d6_call, 'UND')
        cyp2c9_activity = d9_activity_dict.get(cyp2c9_call, 'UND')

        # Use the comparison results to populate the Metabolizer Status and Activity Score columns of the allele DataFrame
        allele_df.loc[index, 'Metabolizer Status'] = f'2D6: {cyp2d6_status} ||| 2C9: {cyp2c9_status}'
        allele_df.loc[index, 'Activity Score'] = f'{cyp2d6_activity}, {cyp2c9_activity}'

    # convert all of the updated data back into a list of lists
    updated_data = allele_df.values.tolist()
    # save_loc(allele_df)
    print(updated_data)
    window['-TABLE-'].update(values=updated_data)

    csv_window(updated_data)

问题分析与解决思路 #

核心原因 #

PySimpleGUI的Table组件需要同时更新**行数据（values）和表头列名（headings）**才能完整展示新增列。当前代码仅更新了行数据，未同步更新表头，导致新增列的名称无法显示。

具体修复步骤 #

1. 获取更新后的列名列表
   在生成updated_data后，从处理后的DataFrame中提取最新列名：

   updated_headings = allele_df.columns.tolist()
   

2. 更新现有Table组件的表头和数据
   修改window['-TABLE-'].update()的调用，同时传入headings参数：

   window['-TABLE-'].update(values=updated_data, headings=updated_headings)
   

3. 修改csv_window函数的调用与实现

   • 调用时传入新列名：csv_window(updated_data, updated_headings)
   • 确保csv_window函数内部在创建或更新Table时，使用传入的updated_headings而非初始列名。例如，若csv_window是新建窗口，需在sg.Table的参数中指定headings=updated_headings；若更新现有窗口，需执行类似table_element.update(headings=updated_headings)的操作。

额外优化建议 #

• 代码中merged_df对象创建后未被使用，可删除以减少冗余；
• 使用iterrows()循环逐行赋值效率较低，建议改用pandas的向量操作或apply方法优化，例如：

  allele_df['Metabolizer Status'] = allele_df.apply(
      lambda row: f"2D6: {d6_status_dict.get(row['CYP2D6'], 'UND')} ||| 2C9: {d9_status_dict.get(row['CYP2C9'], 'UND')}",
      axis=1
  )
  allele_df['Activity Score'] = allele_df.apply(
      lambda row: f"{d6_activity_dict.get(row['CYP2D6'], 'UND')}, {d9_activity_dict.get(row['CYP2C9'], 'UND')}",
      axis=1
  )
  

内容的提问来源于stack exchange，提问作者ClarkThark