根据最近的 DateTime 合并两个 Pandas DataFrame

在本文中，我们将讨论如何根据最近的 DateTime 合并 Pandas DataFrame。要首先了解如何合并 DataFrame，您必须了解如何为其创建 DataFrame，您必须参考文章创建 Pandas DataFrame 。创建 DataFrames 后需要合并它们并合并 Dataframe 有一个名为merge_asof()的函数在编写时可以写为：

pandas.merge_asof(left, right, on=None, left_on=None, right_on=None, left_index=False, right_index=False, by=None, left_by=None, right_by=None, suffixes=(‘_x’, ‘_y’), tolerance=None, allow_exact_matches=True, direction=’backward’)

笔记：

•  要了解有关此函数的更多信息，请参阅文章 Python的pandas.merge_asof()函数
• 数据帧必须按键排序。

循序渐进的方法

第一步：导入pandas库

要完成此任务，我们必须导入名为 Pandas 的库。

import pandas as pd

第 2 步：创建数据框

在这一步中，我们必须使用函数“pd.DataFrame()”创建数据帧。在此，我们创建了 2 个数据帧，一个名为 left，另一个名为 right，因为我们的最后目标是基于最近的 DataTime 合并 2 个数据帧。可以写成：

left = pd.DataFrame( {

       “time”: [pd.Timestamp(“2020-03-25 13:30:00.023”),

           pd.Timestamp(“2020-03-25 13:30:00.023”),

           pd.Timestamp(“2020-03-25 13:30:00.030”),

           pd.Timestamp(“2020-03-25 13:30:00.041”),

           pd.Timestamp(“2020-03-25 13:30:00.048”),

           pd.Timestamp(“2020-03-25 13:30:00.049”),

           pd.Timestamp(“2020-03-25 13:30:00.072”),

           pd.Timestamp(“2020-03-25 13:30:00.075”)

       ],

       “ticker”: [“GOOG”,”MSFT”,”MSFT”,”MSFT”,”GOOG”,”AAPL”,”GOOG”,”MSFT”],

          “bid”: [720.50, 51.95, 51.97, 51.99, 720.50, 97.99, 720.50, 52.01],

          “ask”: [720.93, 51.96, 51.98, 52.00, 720.93, 98.01, 720.88, 52.03]

   })

right = pd.DataFrame( {

          “time”: [

              pd.Timestamp(“2020-03-25 13:30:00.023”),

              pd.Timestamp(“2020-03-25 13:30:00.038”),

              pd.Timestamp(“2020-03-25 13:30:00.048”),

              pd.Timestamp(“2020-03-25 13:30:00.048”),

              pd.Timestamp(“2020-03-25 13:30:00.048”)

          ],

          “ticker”: [“MSFT”, “MSFT”, “GOOG”, “GOOG”, “AAPL”],

          “price”: [51.95, 51.95, 720.77, 720.92, 98.0],

          “quantity”: [75, 155, 100, 100, 100]

      })

第 3 步：合并数据框并打印它们

在这一步中，将使用函数“pd.merge_asof()”合并数据帧。 merge_asof()函数的结果存储在一个变量中，然后使用“print()”打印该变量。

Python3

# Importing the required package
import pandas as pd
  
# Creating the DataFrame of left side
left = pd.DataFrame({
    
    "time": [pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.030"),
             pd.Timestamp("2020-03-25 13:30:00.041"),
             pd.Timestamp("2020-03-25 13:30:00.048"),
             pd.Timestamp("2020-03-25 13:30:00.049"),
             pd.Timestamp("2020-03-25 13:30:00.072"),
             pd.Timestamp("2020-03-25 13:30:00.075")
             ],
    
    "ticker": ["GOOG", "MSFT", "MSFT", "MSFT", "GOOG",
               "AAPL", "GOOG", "MSFT"],
    
    "bid": [720.50, 51.95, 51.97, 51.99, 720.50, 97.99, 
            720.50, 52.01],
    
    "ask": [720.93, 51.96, 51.98, 52.00, 720.93, 98.01, 
            720.88, 52.03]
})
  
# Creating the Dataframe of right side
right = pd.DataFrame({
    "time": [
        pd.Timestamp("2020-03-25 13:30:00.023"),
        pd.Timestamp("2020-03-25 13:30:00.038"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048")
    ],
    "ticker": ["MSFT", "MSFT", "GOOG", "GOOG", "AAPL"],
    
    "price": [51.95, 51.95, 720.77, 720.92, 98.0],
    
    "quantity": [75, 155, 100, 100, 100]
})
  
# Applying merge_asof on data and store it
# in a variable
merged_dataframe = pd.merge_asof(right, left, on="time",
                                 by="ticker")
  
# print the variable
print(merged_dataframe)

输出 ：

示例 1：现在我们在 merge_asof函数更改左右 Dataframe 的位置。

蟒蛇3

# Importing the required package
import pandas as pd
# Creating the DataFrame of left side
left = pd.DataFrame({
    "time": [pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.030"),
             pd.Timestamp("2020-03-25 13:30:00.041"),
             pd.Timestamp("2020-03-25 13:30:00.048"),
             pd.Timestamp("2020-03-25 13:30:00.049"),
             pd.Timestamp("2020-03-25 13:30:00.072"),
             pd.Timestamp("2020-03-25 13:30:00.075")
             ],
    "ticker": ["GOOG", "MSFT", "MSFT", "MSFT", "GOOG",
               "AAPL", "GOOG", "MSFT"],
    
    "bid": [720.50, 51.95, 51.97, 51.99, 720.50, 97.99,
            720.50, 52.01],
    
    "ask": [720.93, 51.96, 51.98, 52.00, 720.93, 98.01,
            720.88, 52.03]
})
  
# Creating the Dataframe of right side
right = pd.DataFrame({
    "time": [
        pd.Timestamp("2020-03-25 13:30:00.023"),
        pd.Timestamp("2020-03-25 13:30:00.038"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048")
    ],
    "ticker": ["MSFT", "MSFT", "GOOG", "GOOG", "AAPL"],
    
    "price": [51.95, 51.95, 720.77, 720.92, 98.0],
    
    "quantity": [75, 155, 100, 100, 100]
})
  
# Applying merge_asof on data and store it 
# in a variable
merged_dataframe = pd.merge_asof(left, right, on="time",
                                 by="ticker")
  
# print the variable
print(merged_dataframe)

输出：

注意：因此，从我们的 2 个输出中可以清楚地看出，当我们将右 DataFrame 放在第一位时，输出中的行数为 5 等于右 DataFrame 中的行数，而当左 DataFrame 放在第一位时那么输出中的行数等于左侧 DataFrame 中的行数。如果我们查看两个输出并比较它们，那么我们可以很容易地说 merge_asof() 类似于左连接，除了我们匹配最近的键而不是相等的键。

示例2：我们只在报价时间和交易时间之间的2ms内进行。

蟒蛇3

# Importing the required package
import pandas as pd
  
# Creating the DataFrame of left side
left = pd.DataFrame({
    "time": [pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.030"),
             pd.Timestamp("2020-03-25 13:30:00.041"),
             pd.Timestamp("2020-03-25 13:30:00.048"),
             pd.Timestamp("2020-03-25 13:30:00.049"),
             pd.Timestamp("2020-03-25 13:30:00.072"),
             pd.Timestamp("2020-03-25 13:30:00.075")
             ],
    "ticker": ["GOOG", "MSFT", "MSFT", "MSFT", "GOOG",
               "AAPL", "GOOG", "MSFT"],
    
    "bid": [720.50, 51.95, 51.97, 51.99, 720.50, 97.99,
            720.50, 52.01],
    
    "ask": [720.93, 51.96, 51.98, 52.00, 720.93, 98.01, 
            720.88, 52.03]
})
  
# Creating the Dataframe of right side
right = pd.DataFrame({
    "time": [
        pd.Timestamp("2020-03-25 13:30:00.023"),
        pd.Timestamp("2020-03-25 13:30:00.038"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048")
    ],
    "ticker": ["MSFT", "MSFT", "GOOG", "GOOG", "AAPL"],
    
    "price": [51.95, 51.95, 720.77, 720.92, 98.0],
    
    "quantity": [75, 155, 100, 100, 100]
})
  
# Applying merge_asof on data and store it
# in a variable
merged_dataframe = pd.merge_asof(left, right, on="time", by="ticker",
                                 tolerance=pd.Timedelta("2ms"))
  
# print the variable
print(merged_dataframe)

输出 ：

示例3：我们只在报价时间和交易时间之间的10ms内进行asof，并且我们排除了时间上的完全匹配。但是，先前的数据将向前传播。

蟒蛇3

# Importing the required package
import pandas as pd
  
# Creating the DataFrame of left side
left = pd.DataFrame({
    "time": [pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.030"),
             pd.Timestamp("2020-03-25 13:30:00.041"),
             pd.Timestamp("2020-03-25 13:30:00.048"),
             pd.Timestamp("2020-03-25 13:30:00.049"),
             pd.Timestamp("2020-03-25 13:30:00.072"),
             pd.Timestamp("2020-03-25 13:30:00.075")
             ],
    "ticker": ["GOOG", "MSFT", "MSFT", "MSFT", "GOOG",
               "AAPL", "GOOG", "MSFT"],
    
    "bid": [720.50, 51.95, 51.97, 51.99, 720.50, 97.99,
            720.50, 52.01],
    
    "ask": [720.93, 51.96, 51.98, 52.00, 720.93, 98.01, 
            720.88, 52.03]
})
  
# Creating the Dataframe of right side
right = pd.DataFrame({
    "time": [
        pd.Timestamp("2020-03-25 13:30:00.023"),
        pd.Timestamp("2020-03-25 13:30:00.038"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048")
    ],
    
    "ticker": ["MSFT", "MSFT", "GOOG", "GOOG", "AAPL"],
    
    "price": [51.95, 51.95, 720.77, 720.92, 98.0],
    
    "quantity": [75, 155, 100, 100, 100]
})
  
# Applying merge_asof on data and store it
# in a variable
merged_dataframe = pd.merge_asof(left, right, on="time", by="ticker",
                                 tolerance=pd.Timedelta("2ms"),
                                 allow_exact_matches=False)
  
# print the variable
print(merged_dataframe)

输出 ：

示例 4：当两个地方都使用相同的 DataFrame 时。在这个左边的 Dataframe 中，两边都使用了。

蟒蛇3

# Importing the required package
import pandas as pd
  
# Creating the DataFrame of left side
left = pd.DataFrame({
    "time": [pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.023"),
             pd.Timestamp("2020-03-25 13:30:00.030"),
             pd.Timestamp("2020-03-25 13:30:00.041"),
             pd.Timestamp("2020-03-25 13:30:00.048"),
             pd.Timestamp("2020-03-25 13:30:00.049"),
             pd.Timestamp("2020-03-25 13:30:00.072"),
             pd.Timestamp("2020-03-25 13:30:00.075")
             ],
    
    "ticker": ["GOOG", "MSFT", "MSFT", "MSFT", "GOOG",
               "AAPL", "GOOG", "MSFT"],
    
    "bid": [720.50, 51.95, 51.97, 51.99, 720.50, 97.99,
            720.50, 52.01],
    
    "ask": [720.93, 51.96, 51.98, 52.00, 720.93, 98.01, 
            720.88, 52.03]
})
  
# Creating the Dataframe of right side
right = pd.DataFrame({
    "time": [
        pd.Timestamp("2020-03-25 13:30:00.023"),
        pd.Timestamp("2020-03-25 13:30:00.038"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048"),
        pd.Timestamp("2020-03-25 13:30:00.048")
    ],
    
    "ticker": ["MSFT", "MSFT", "GOOG", "GOOG", "AAPL"],
    
    "price": [51.95, 51.95, 720.77, 720.92, 98.0],
    
    "quantity": [75, 155, 100, 100, 100]
})
  
# Applying merge_asof on data and store it
# in a variable
merged_dataframe = pd.merge_asof(left, left, on="time", 
                                 by="ticker")
  
# print the variable
print(merged_dataframe)

输出 ：

它将相同的数据帧创建为 2 个帧，一个表示为 x，另一个表示为 y，即bid_x、bid_y、ask_x、ask_y。