我有以下csv数据，public class Controller
{
public HeaderObject HeaderObj;
public Main_Form FormMain;
Public Controller(Main_Form pFormMain)
{
FormMain = pFormMain;
}
......
foreach(var sheet in ExcelSheets)
{
PrintToConsole(sheet + " - Processing Data");
if (formPreferences.CheckEverySheet.Checked)
{
HeaderObj = null; // set to null each time we process a new sheet
FormMain.formSheetChecker = new Form_Headers(lFormMain, this); //pass a reference to Controller class here
QuickLoadExcelData(lExcelWorkSheet);
FormMain.formSheetChecker.StartPosition = FormStartPosition.CenterScreen;
FormMain.formSheetChecker.ShowDialog(); //do some processing here and then pass back object to controller class before form
//before the formSheetChecker is closed it sets the HeaderObj in the controller class as i have passed a reference to the controller class in the
//formSheetCheckers constructor.
if (HeaderObj != null)
{
storageObject.setHeader(lExcelWorkSheet.Name, HeaderObj); //i get a NullReferenceException here even when there is data inside the object
}
else
{
HeaderObj = new HeaderObject("failed", -1);
storageObject.setHeader(lExcelWorkSheet.Name, HeaderObj);
}
}
}
}
public partial class Form_Headers : Form
{
Form_Main FormMain;
Controller controller;
HeaderObject headerObject = null;
string lSheetName = "";
public Form_Headers(Form_Main pFormMain, Controller pController) {
InitializeComponent();
FormMain = pFormMain;
controller = pController;
headerObject = controller.HeaderObj;
}
public void UpdateHeaderRange() {
if (!(headerObject.Variable1 == ""))
{
//do nothing here
}
else {
headerObject.SetHeaderFields("failed", -1);
}
}
private void button_Save_Click(object sender, EventArgs e)
{
UpdateHeaderRange();
this.Close();
}
。请注意，csv文件格式正确，并且不包含任何“怪异”的东西。
data.csv我通过以下方式阅读它：
id,date,time,name,value,description
0,2017-01-07,2017-01-07 11:49:29,cool_name1,2,something
1,2017-01-07,2017-01-07 11:51:50,cool_name1,-1,something
2,2017-01-07,2017-01-07 14:03:42,cool_name2,9,something
结果为schema = {'id': float,
'date': str,
'time': str,
'name': str,
'value': float,
'description': str}
df = pd.read_csv('data.csv', header=0, dtype=schema,
parse_dates=['date', 'time'])
：
df.dtypes由于数据的去向，id int64
date datetime64[ns]
time datetime64[ns]
name object
value float64
description object
必须按照id（在这种情况下为schema）的指定来自动解析。但是，在尝试了一堆不同的关键字参数并相互结合之后，我无法让float服从我的id。
我尝试过的内容的列表：
dtype应注意，列dtype=schema
converters=schema
low_memory=False
engine='c' AND 'python'
index_col=False
parse_dates=False
确实服从value，并且没有指定它就变成dtype。
缺少的成分是什么（熊猫版0.23.4）？