循环遍历df的行，并根据IFF语句执行计算

Question

我正试图用几种货币将我的经纪费换算成美元。因此，如果我以欧元支付费用，费用应乘以该天的欧元/美元汇率

数据帧示例：

import pandas as pd
df = pd.DataFrame({"Fee Coin": ["EUR", "BTC"],
                   "Fee": [3, 0.0005],
                    "USD Price": [1.05, 1.1],
                     "BTC Price": [24000, 27000]})

第一行的输出应为3.15，第二行的输出应为13.5。输出应显示在一个新的列中，我已尝试调用费用（美元）

我所尝试的，给出了错误的计算：

for value in df["Fee Coin"]:
    if value == "EUR":
        df["Fee in USD"] = df.Fee*df["USD Price"]
    elif value == "BTC":
        df["Fee in USD"] = df.Fee*df["BTC Price"]

Answer 1

因为您只有两个类别：EUR和BTC，所以可以使用np.where

df["Fee in USD"] = np.where(df["Fee Coin"] == "EUR", df.Fee * df["USD Price"], 
                             df.Fee * df["BTC Price"])

输出：

   Fee Coin     Fee  USD Price  BTC Price  Fee in USD
0      EUR  3.0000       1.05      24000        3.15
1      BTC  0.0005       1.10      27000       13.50

更新： 对于多种货币，最好有一张从收费硬币到列名的地图，即

currency_to_col = {"EUR": "USD Price", "BTC": "BTC Price", "JPY": "JPY Price"}.

然后使用np.select

condlist = [df["Fee Coin"] == currency for currency in df["Fee Coin"].unique()]
choicelist = [df.Fee * df[currency_to_col[currency]] for currency in df["Fee Coin"].unique()]
df["Fee in USD"] = np.select(condlist, choicelist)

数据帧：

  Fee Coin     Fee  USD Price  BTC Price  JPY Price
0      EUR  3.0000       1.05      24000       1000
1      BTC  0.0005       1.10      27000       2000
2      JPY  0.2000       1.20      29000       3000

输出：

  Fee Coin     Fee  USD Price  BTC Price  JPY Price  Fee in USD
0      EUR  3.0000       1.05      24000       1000        3.15
1      BTC  0.0005       1.10      27000       2000       13.50
2      JPY  0.2000       1.20      29000       3000      600.00

Answer 2

m_eur = df["Fee Coin"] == "EUR"
m_btc = df["Fee Coin"] == "BTC"

df["Fee in USD"] = pd.concat(
    [
        df.loc[m_eur, "Fee"] * df.loc[m_eur, "USD Price"],
        df.loc[m_btc, "Fee"] * df.loc[m_btc, "BTC Price"],
    ]
)

print(df)

印刷品：

  Fee Coin     Fee  USD Price  BTC Price  Fee in USD
0      EUR  3.0000       1.05      24000        3.15
1      BTC  0.0005       1.10      27000       13.50

Answer 3

计算错误的原因是，对于每一行，都要乘以整个df['Fee'] column with df['USD Price'] or df['BTC Price']。基本上基于 在最后一行的费用硬币价值上，您的费用栏将被USD/BTC价格栏乘以

在您的情况下，由于最后一行包含“BTC”硬币，因此它将乘以BTC价格，结果如下：

   Fee Coin  Fee     USD Price  BTC Price  Fee in USD
0  EUR       3       1.05       24000      72000
1  BTC       0.0005  1.1        27000      13.5

因此，基本上应该乘以特定的行值，而不是乘以整个列

for i in range(len(df)):
    if df.loc[i, "Fee Coin"] == "EUR":
        df.loc[i,"Fee in USD"] = df.loc[i,"Fee"]*df.loc[i,"USD Price"]
    elif df.loc[i, "Fee Coin"] == "BTC":
        df.loc[i,"Fee in USD"] = df.loc[i,"Fee"]*df.loc[i,"BTC Price"]

最终数据帧：

   Fee Coin  Fee     USD Price  BTC Price  Fee in USD
0  EUR       3       1.05       24000      3.15
1  BTC       0.0005  1.1        27000      13.50

p.S.对于三种或三种以上的货币，您可以在条件If Else块中添加条件