如何解决熊猫-每个周期后重置CUMSUM
我打算搬到公司的另一个职位，我要求进行典型的任务来训练自己。在某种程度上，我做到了，我真的很了解，但是现在我被困住了。我当时在搜索，但是对我来说还算不上什么，因为我也不明白，但是我想那应该是某种循环，每次使用循环时都会出错。
基本上，我有一个巨大的Excel文件，其中包含列和大量数据。它是关于出售食物和客户费用的，关于收费的错误，后来又被修正了，我不得不弄清楚余额到底在哪里得到0.00。
在我开始处理之前，您可以在这里看到数据。
Original Data,before I used the pivot table
后来，我使用数据透视表将行与一列分开，并发现了将来可能出现的不同数据和问题。因此，这是我处理后的excel文件。
Data after working on it,with additional columns
这是我当前的代码：
import numpy as np
import pandas as pd
import xlrd
from pandas import Series,DataFrame
df = pd.read_excel ('C:/Data.xlsx',sheet_name = 'Sheet1',usecols = ['Payment','Money','Created'])
df['Created'] = pd.to_datetime(df['Created'])
df['Created'] = df['Created'].dt.round('min')
df = df.pivot_table(index = ['Created'],columns = ['Payment']).fillna(0)
df['Money','fee'] = df['Money','fee'].round(2)
df['Fixed Fee'] = (-df['Money','food'] * 25) / 100
df['Fixed Fee'] = df['Fixed Fee'].round(2)
df['OverCharge'] = np.where(df['Money','fee'] != df['Fixed Fee'],df['Money','fee'] - df['Fixed Fee'],0)
df['OverCharge'] = df['OverCharge'].round(1)
df['Percentage'] = df['Money','fee'] / df['Money','food'] * 100
df['Percentage'] = df['Percentage'].abs()
df['Percentage'] = df['Percentage'].round(2)
df['Charges'] = np.where(df['Percentage'].notna(),np.where(df['Percentage'] > 26,'Overcharge - 30%','Fixed - 25%'),'Null')
df['Correct'] = -df['Money','food'] - df['Fixed Fee']
df['Incorrect'] = -df['Money','food'] - df['Money','fee']
df['Balance'] = df['Correct'] - df['Incorrect']
df['Balance'] = np.where(df['Money','payout'] != 0,df['Correct'].cumsum() - df['Money','payout'],df['Balance'])
#df.to_excel("CarIndustry.xlsx")
print(df)
输出（前20行）：
Money Fixed Fee OverCharge Percentage Charges Correct Incorrect Balance
Payment fee food payout payoutReject
Created
2019-12-27 12:32:00 -6.00 20.0 0.00 0.0 -5.00 -1.0 30.00 Overcharge - 30% -15.00 -14.00 -1.00
2019-12-27 12:58:00 -5.26 17.5 0.00 0.0 -4.38 -0.9 30.06 Overcharge - 30% -13.12 -12.24 -0.88
2019-12-27 13:17:00 -3.46 11.5 0.00 0.0 -2.88 -0.6 30.09 Overcharge - 30% -8.62 -8.04 -0.58
2019-12-30 04:01:00 0.00 0.0 -34.29 0.0 -0.00 0.0 NaN Null 0.00 -0.00 -2.45
2019-12-30 13:24:00 -1.94 6.5 0.00 0.0 -1.62 -0.3 29.85 Overcharge - 30% -4.88 -4.56 -0.32
2020-01-01 12:53:00 -6.00 20.0 0.00 0.0 -5.00 -1.0 30.00 Overcharge - 30% -15.00 -14.00 -1.00
2020-01-01 13:06:00 -3.90 13.0 0.00 0.0 -3.25 -0.6 30.00 Overcharge - 30% -9.75 -9.10 -0.65
2020-01-01 13:27:00 -3.46 11.5 0.00 0.0 -2.88 -0.6 30.09 Overcharge - 30% -8.62 -8.04 -0.58
2020-01-01 13:38:00 -7.20 24.0 0.00 0.0 -6.00 -1.2 30.00 Overcharge - 30% -18.00 -16.80 -1.20
2020-01-01 15:10:00 -2.10 7.0 0.00 0.0 -1.75 -0.4 30.00 Overcharge - 30% -5.25 -4.90 -0.35
2020-01-01 16:31:00 -7.94 26.5 0.00 0.0 -6.62 -1.3 29.96 Overcharge - 30% -19.88 -18.56 -1.32
2020-01-01 16:51:00 -2.40 8.0 0.00 0.0 -2.00 -0.4 30.00 Overcharge - 30% -6.00 -5.60 -0.40
2020-01-01 17:00:00 -2.26 7.5 0.00 0.0 -1.88 -0.4 30.13 Overcharge - 30% -5.62 -5.24 -0.38
2020-01-01 18:21:00 -8.26 27.5 0.00 0.0 -6.88 -1.4 30.04 Overcharge - 30% -20.62 -19.24 -1.38
2020-01-03 13:24:00 -1.66 5.5 0.00 0.0 -1.38 -0.3 30.18 Overcharge - 30% -4.12 -3.84 -0.28
2020-01-03 15:53:00 -3.30 11.0 0.00 0.0 -2.75 -0.5 30.00 Overcharge - 30% -8.25 -7.70 -0.55
2020-01-03 17:39:00 -1.94 6.5 0.00 0.0 -1.62 -0.3 29.85 Overcharge - 30% -4.88 -4.56 -0.32
2020-01-03 20:22:00 -3.14 10.5 0.00 0.0 -2.62 -0.5 29.90 Overcharge - 30% -7.88 -7.36 -0.52
2020-01-03 21:18:00 -2.26 7.5 0.00 0.0 -1.88 -0.4 30.13 Overcharge - 30% -5.62 -5.24 -0.38
2020-01-06 04:01:00 0.00 0.0 -134.75 0.0 -0.00 0.0 NaN Null 0.00 -0.00 -46.36
我的结果应该是：
Money Fixed Fee OverCharge Percentage Charges Correct Incorrect Balance
Payment fee food payout payoutReject
Created
2019-12-27 12:32:00 -6.00 20.0 0.00 0.0 -5.00 -1.0 30.00 Overcharge - 30% -15.00 -14.00 -1.00
2019-12-27 12:58:00 -5.26 17.5 0.00 0.0 -4.38 -0.9 30.06 Overcharge - 30% -13.12 -12.24 -0.88
2019-12-27 13:17:00 -3.46 11.5 0.00 0.0 -2.88 -0.6 30.09 Overcharge - 30% -8.62 -8.04 -0.58
2019-12-30 04:01:00 0.00 0.0 -34.29 0.0 -0.00 0.0 NaN Null 0.00 -0.00 -2.45
2019-12-30 13:24:00 -1.94 6.5 0.00 0.0 -1.62 -0.3 29.85 Overcharge - 30% -4.88 -4.56 -0.32
2020-01-01 12:53:00 -6.00 20.0 0.00 0.0 -5.00 -1.0 30.00 Overcharge - 30% -15.00 -14.00 -1.00
2020-01-01 13:06:00 -3.90 13.0 0.00 0.0 -3.25 -0.6 30.00 Overcharge - 30% -9.75 -9.10 -0.65
2020-01-01 13:27:00 -3.46 11.5 0.00 0.0 -2.88 -0.6 30.09 Overcharge - 30% -8.62 -8.04 -0.58
2020-01-01 13:38:00 -7.20 24.0 0.00 0.0 -6.00 -1.2 30.00 Overcharge - 30% -18.00 -16.80 -1.20
2020-01-01 15:10:00 -2.10 7.0 0.00 0.0 -1.75 -0.4 30.00 Overcharge - 30% -5.25 -4.90 -0.35
2020-01-01 16:31:00 -7.94 26.5 0.00 0.0 -6.62 -1.3 29.96 Overcharge - 30% -19.88 -18.56 -1.32
2020-01-01 16:51:00 -2.40 8.0 0.00 0.0 -2.00 -0.4 30.00 Overcharge - 30% -6.00 -5.60 -0.40
2020-01-01 17:00:00 -2.26 7.5 0.00 0.0 -1.88 -0.4 30.13 Overcharge - 30% -5.62 -5.24 -0.38
2020-01-01 18:21:00 -8.26 27.5 0.00 0.0 -6.88 -1.4 30.04 Overcharge - 30% -20.62 -19.24 -1.38
2020-01-03 13:24:00 -1.66 5.5 0.00 0.0 -1.38 -0.3 30.18 Overcharge - 30% -4.12 -3.84 -0.28
2020-01-03 15:53:00 -3.30 11.0 0.00 0.0 -2.75 -0.5 30.00 Overcharge - 30% -8.25 -7.70 -0.55
2020-01-03 17:39:00 -1.94 6.5 0.00 0.0 -1.62 -0.3 29.85 Overcharge - 30% -4.88 -4.56 -0.32
2020-01-03 20:22:00 -3.14 10.5 0.00 0.0 -2.62 -0.5 29.90 Overcharge - 30% -7.88 -7.36 -0.52
2020-01-03 21:18:00 -2.26 7.5 0.00 0.0 -1.88 -0.4 30.13 Overcharge - 30% -5.62 -5.24 -0.38
2020-01-06 04:01:00 0.00 0.0 -134.75 0.0 -0.00 0.0 NaN Null 0.00 -0.00 -12.07
要更好地理解它，只需检查最后几行，基本上我想让cumsum从上一个支出到下一个支出进行计算，然后将此数字放在最后，然后像这样继续，因为这只是其中的一小部分数据。
解决方法
因此，最简单的方法是更改余额列的计算方式，以使其首次实现该目的。
class SchedulePage extends Component {
constructor(props) {
super(props);
this.state = {
events: [
{
start: start_date,end: end_date,title: "15 minute interview",},],};
}
render() {
if (!this.state.interview_dates) { // Updated to cover null,undefined,false
return null;
}
const data = this.state.interview_dates.map((item) => {
return {
start: moment(item.start).toDate(),end: moment(item.end).toDate(),title: item.title,};
});
console.log(data);
return (
<div className="calendar-container">
<Calendar
localizer={localizer}
views={["month","week"]}
defaultDate={new Date()}
defaultView="week"
events={data}
style={{ height: "100vh" }}
onSelectEvent={this.test}
step={15}
//min={new Date(2020,1,6,0)}
//max={new Date(2020,23,0)}
/>
<ScheduleInterview
onClose={this.showCreateInterview}
show={this.state.createInterview}
/>
</div>
);
}
因此，我们在每个组上都使用cumsum，其中每个组都定义为直到“正确”列值等于0为止的所有行。
这将替换您的df ['balance']行。
版权声明：本文内容由互联网用户自发贡献，该文观点与技术仅代表作者本人。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容， 请发送邮件至 dio@foxmail.com 举报，一经查实，本站将立刻删除。
