我想提取一个单词后，一个固定的组合词和特殊字符，也需要计数提取的单词在特定的单元格。

Question

我有一个数据帧df，它有一个列。我想在一个固定的词组合后提取一个单词&特殊字符，也需要一个特定单元格中提取的单词数。

例如：(最近的警报触发器)“、”valueString“：”倾斜传感器“、”最新的警报触发器“、”valueString“：”Hello world“、(最新的警报触发器)、”valueString“：”ABC“，

现在，从上面的行中，我想在“(最近的警报触发器)”、“valueString”之后在逗号之间提取任何单词：

因此，在这种情况下，我只想要‘倾斜传感器’和它的计数在特定的细胞。。

我不需要‘你好世界’或'ABC‘，因为它是第二或第三。基本上我要的是第一个搜索词。

以下是我的df:-

import re
import pandas as pd
import numpy as np



data = {'product_name': ["[{'name':'Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'Tilt Sensor','packetType':'enumerated','leastSigBit':440,,'Tilt Sensor','mostSigBit':447},{'name':'User Set Year (Most Recent Alarm Trigger)', (Most Recent Alarm Trigger)','valueString':'Band','valueNumber':2022.0,'units':'Undefined / Not Used','packetType':'Tilt Sensor','leastSigBit':432,'mostSigBit':439},{'name':'User Set Month,(Most Recent Alarm Trigger)','valueString':'Back space',{'name':'User Set Minute (Most Recent Alarm Trigger)','valueNumber':16.0,'units':'min','packetType':'unsigned','leastSigBit':400,'mostSigBit':407},'Tilt Sensor',{'name':'User Set Second (Most Recent Alarm Trigger)','valueNumber':36.0,'units':'s','packetType':'unsigned','leastSigBit':392,'mostSigBit':399}]",
                         "[{'name':'Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'Volumetric Sensor','packetType':'enumerated','leastSigBit':440,'mostSigBit':447},{'name':'User Set Year (Most Recent Alarm Trigger)','valueNumber':2022.0,'units':'(Most Recent Alarm Trigger)','valueString':'Being human','packetType':'unsigned','leastSigBit':432,'mostSigBit':439},{'name':'User Set Month (Most Recent Alarm Trigger)','valueNumber':6.0,'(Most Recent Alarm Trigger)','valueString':'Hello'':'Month','Volumetric Sensor','packetType':'unsigned','leastSigBit':424,'mostSigBit':431},{'name':'User Set Day (Most Recent Alarm ]"]}
df = pd.DataFrame(data)
df

我尝试了regex或应用方法，但没有得到我想要的。

下面是一些我尝试过的代码，

df["Extract"] = df["product_name"].apply(lambda st: st[st.find("(Most Recent Alarm Trigger)','valueString':")+1:st.find(",")])

df['Title'] = df.product_name.str.extract(r'"(Most Recent Alarm Trigger)','valueString':'"\s*([^\.]*)\s*\.', expand=False)

以下是我的预期结果：

data = {'product_name': ["[{'name':'Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'Tilt Sensor','packetType':'enumerated','leastSigBit':440,,'Tilt Sensor','mostSigBit':447},{'name':'User Set Year (Most Recent Alarm Trigger)', (Most Recent Alarm Trigger)','valueString':'Band','valueNumber':2022.0,'units':'Undefined / Not Used','packetType':'Tilt Sensor','leastSigBit':432,'mostSigBit':439},{'name':'User Set Month,(Most Recent Alarm Trigger)','valueString':'Back space',{'name':'User Set Minute (Most Recent Alarm Trigger)','valueNumber':16.0,'units':'min','packetType':'unsigned','leastSigBit':400,'mostSigBit':407},'Tilt Sensor',{'name':'User Set Second (Most Recent Alarm Trigger)','valueNumber':36.0,'units':'s','packetType':'unsigned','leastSigBit':392,'mostSigBit':399}]",
                             "[{'name':'Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'Volumetric Sensor','packetType':'enumerated','leastSigBit':440,'mostSigBit':447},{'name':'User Set Year (Most Recent Alarm Trigger)','valueNumber':2022.0,'units':'(Most Recent Alarm Trigger)','valueString':'Being human','packetType':'unsigned','leastSigBit':432,'mostSigBit':439},{'name':'User Set Month (Most Recent Alarm Trigger)','valueNumber':6.0,'(Most Recent Alarm Trigger)','valueString':'Hello'':'Month','Volumetric Sensor','packetType':'unsigned','leastSigBit':424,'mostSigBit':431},{'name':'User Set Day (Most Recent Alarm ]"],
       'Extarct': ['Tilt Sensor','Volumetric Sensor'],'Count': [4,2]}
df = pd.DataFrame(data)
df

Answer 1

一种解决办法可以如下：

\',.

• Next，
• 使用Series.str.extract来获得\'valueString\':\'和df.apply之间的第一次匹配--使用df.apply为每一行(axis=1)提供一个lambda函数，以获得现在存储在适当product_name字符串中的df.Extract中的每个值的计数。

import pandas as pd

# also adding the string from your comment
data = {'product_name': ["[{'name':'Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'Tilt Sensor','packetType':'enumerated','leastSigBit':440,,'Tilt Sensor','mostSigBit':447},{'name':'User Set Year (Most Recent Alarm Trigger)', (Most Recent Alarm Trigger)','valueString':'Band','valueNumber':2022.0,'units':'Undefined / Not Used','packetType':'Tilt Sensor','leastSigBit':432,'mostSigBit':439},{'name':'User Set Month,(Most Recent Alarm Trigger)','valueString':'Back space',{'name':'User Set Minute (Most Recent Alarm Trigger)','valueNumber':16.0,'units':'min','packetType':'unsigned','leastSigBit':400,'mostSigBit':407},'Tilt Sensor',{'name':'User Set Second (Most Recent Alarm Trigger)','valueNumber':36.0,'units':'s','packetType':'unsigned','leastSigBit':392,'mostSigBit':399}]",
                         "[{'name':'Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'Volumetric Sensor','packetType':'enumerated','leastSigBit':440,'mostSigBit':447},{'name':'User Set Year (Most Recent Alarm Trigger)','valueNumber':2022.0,'units':'(Most Recent Alarm Trigger)','valueString':'Being human','packetType':'unsigned','leastSigBit':432,'mostSigBit':439},{'name':'User Set Month (Most Recent Alarm Trigger)','valueNumber':6.0,'(Most Recent Alarm Trigger)','valueString':'Hello'':'Month','Volumetric Sensor','packetType':'unsigned','leastSigBit':424,'mostSigBit':431},{'name':'User Set Day (Most Recent Alarm ]",
                         "[{'name':'Power Mode Quality Factor','valueString':'Power Mode Undefined','valueString':'Finally',Trigger Cause Status (Most Recent Alarm Trigger)','valueString':'No Trigger (Event Store Empty)',}]"]}
df = pd.DataFrame(data)

df['Extract'] = df.product_name.str.extract(
    r'\(Most Recent Alarm Trigger\)\',\'valueString\':\'(.*?)\',')
# N.B. We're using the question mark to make the search for '.*' lazy

df['Count'] = df.apply(lambda row: row.product_name.count(row.Extract), axis=1)

print(df.iloc[:,1:])

                          Extract  Count
0                     Tilt Sensor      4
1               Volumetric Sensor      2
2  No Trigger (Event Store Empty)      1

注:如果str.extract无法找到匹配项，您将在df.Extract中得到NaN值。如果是这样的话，这将导致df.apply(lambda row: row.product_name.count(row.Extract), axis=1)的一个错误(因为它期待一个string)。为了避免这种情况，您可以使用：

df['Count'] = df.apply(lambda row: row.product_name.count(row.Extract) 
                       if isinstance(row.Extract,str) else 0, axis=1)