熊猫数据框到csv引号答案

【问题标题】：pandas dataframe to csv quotation mark熊猫数据框到csv引号
【发布时间】：2017-07-10 14:29:58
【问题描述】：

我尝试使用将 pandas 数据帧转换为 csv

dataframe.to_csv('final_processed.csv', encoding='utf-8', index=False)

然后我得到了 csv 文件，它有 5 列，第一列是文本，我打开 csv 文件，发现有些行在第一列的开头和结尾都用引号引起来，而另一些则没有（如下所示）。我希望它们都以引号开头和结尾，我该怎么办？非常感谢。

“Pt 因胸痛来到急诊室...... PT/INR 和 PTT 在 0324 进行检查， PT/INR 为 92 .5/8 .8 和 PTT wa 249.",管理/监控,非处方/订单,管理/监控,notadr

“患者于 2015 年 5 月 4 日在办公室看到...... 另一个看护人并向我的诊所报告经理。”，转录/准备/分配，不开/订单，不管理/监控，notadr

多种药物由不同的提供者输入 Med Rec 剂量不准确和......医院入院，开处方/订单，开处方/订单，不管理/监控，notadr

【问题讨论】：

标签： python csv pandas data-processing

【解决方案1】：

您需要将quoting=True添加到dataframe.to_csv，如下：

dataframe.to_csv('final_processed.csv', encoding='utf-8', index=False, quoting=True)

您可以在官方文档here 中找到更多详细信息。

这是一个完整的示例

import pandas as pd 
from StringIO import StringIO 

st = """
"Pt come to ER with chest pain ..... PT/INR and PTT be check at 0324, PT/INR as 92 .5/8 .8 and PTT wa 249.",administer/monitor,notprescribe/order,administer/monitor,notadr

"Patient wa see in office on 05/04/2015 ....... Thi wa catch by another caregiver and report to my clinic manager.",transcribe/prepare/dispense,notprescribe/order,notadminister/monitor,notadr

Multiple medication enter on Med Rec by different provider with inaccurate dose and ......hospital admission,prescribe/order,prescribe/order,notadminister/monitor,notadr
"""
data =pd.read_csv(StringIO(st), delimiter=",", header=None) 
data.to_csv("final_processed.csv", encoding="utf-8", index=False, quoting=True)

输出如下所示：

 "0","1","2","3","4"
"Pt come to ER with chest pain ..... PT/INR and PTT be check at 0324, PT/INR as 92 .5/8 .8 and PTT wa 249.","administer/monitor","notprescribe/order","administer/monitor","notadr"
"Patient wa see in office on 05/04/2015 ....... Thi wa catch by another caregiver and report to my clinic manager.","transcribe/prepare/dispense","notprescribe/order","notadminister/monitor","notadr"
"Multiple medication enter on Med Rec by different provider with inaccurate dose and ......hospital admission","prescribe/order","prescribe/order","notadminister/monitor","notadr"

【讨论】：

quoting=True：文件“pandas\core\generic.py”，第 3020 行，在 to_csv：formatter.save() 文件“pandas\io\formats\csvs.py”，第 170 行，在保存中：self.writer = UnicodeWriter(f, **writer_kwargs) 文件“pandas\io\common.py”，第 545 行，在 UnicodeWriter 中：返回 csv.writer(f, dialect=dialect, **kwds) TypeError:" quoting" 必须是整数
你的意思是doublequote？