提高转化过程的效率

2024-10-06 16:27:40 发布

您现在位置：Python中文网/ 问答频道 /正文
7069 3

现在，我有一个从URL中提取大量数据（约150万行）的过程，该URL以无组织的方式出现，我需要随后重新组织。当前的进程运行得非常完美，但是内存非常大，效率也非常低，所以我一直在寻求帮助

我收到的数据如下所示：（注意，在出口8后，还有5列Na，没有一列表示当前SCP结束

['C/A','UNIT','SCP','DATE1','TIME1','DESC1','ENTRIES1','EXITS1','DATE2','TIME2','ESC2',\
 'ENTRIES2','EXITS2','DATE3','TIME3','DESC3','ENTRIES3','EXITS3','DATE4','TIME4','DESC4',\
 'ENTRIES4','EXITS4','DATE5','TIME5','DESC5','ENTRIES5','EXITS5','DATE6','TIME6','DESC6',\
 'ENTRIES6','EXITS6','DATE7','TIME7','DESC7','ENTRIES7','EXITS7','DATE8','TIME8','DESC8',\
 'ENTRIES8','EXITS8']

我的目标是对其进行如下重组：

['c/a','unit','scp','date','time','description','entries','exit']

原始输出示例：

     C/A        UNIT    SCP     DATE1   TIME1   DESC1   ENTRIES1    EXITS1  DATE2   TIME2   ESC2    ENTRIES2    EXITS2  DATE3   TIME3   DESC3   ENTRIES3    EXITS3  DATE4   TIME4   DESC4   ENTRIES4    EXITS4  DATE5   TIME5   DESC5   ENTRIES5    EXITS5  DATE6   TIME6   DESC6   ENTRIES6    EXITS6  DATE7   TIME7   DESC7   ENTRIES7    EXITS7  DATE8   TIME8   DESC8   ENTRIES8    EXITS8
0   A002    R051    02-00-00    04-20-13    00:00:00    REGULAR 4084276 1405308 04-20-13    04:00:00    REGULAR 4084308.0   1405312.0   04-20-13    08:00:00    REGULAR 4084332.0   1405348.0   04-20-13    12:00:00    REGULAR 4084429.0   1405441.0   04-20-13    16:00:00    REGULAR 4084714.0   1405494.0   04-20-13    20:00:00    REGULAR 4085107.0   1405550.0   04-21-13    00:00:00    REGULAR 4085286.0   1405578.0   04-21-13    04:00:00    REGULAR 4085317.0   1405582.0
1   A002    R051    02-00-00    04-21-13    08:00:00    REGULAR 4085336 1405603 04-21-13    12:00:00    REGULAR 4085421.0   1405673.0   04-21-13    16:00:00    REGULAR 4085543.0   1405725.0   04-21-13    20:00:00    REGULAR 4085543.0   1405781.0   04-22-13    00:00:00    REGULAR 4085669.0   1405820.0   04-22-13    04:00:00    REGULAR 4085684.0   1405825.0   04-22-13    08:00:00    REGULAR 4085715.0   1405929.0   04-22-13    12:00:00    REGULAR 4085878.0   1406175.0
2   A002    R051    02-00-00    04-22-13    16:00:00    REGULAR 4086116 1406242 04-22-13    20:00:00    REGULAR 4086986.0   1406310.0   04-23-13    00:00:00    REGULAR 4087164.0   1406335.0   04-23-13    04:00:00    REGULAR 4087172.0   1406339.0   04-23-13    08:00:00    REGULAR 4087214.0   1406441.0   04-23-13    12:00:00    REGULAR 4087390.0   1406685.0   04-23-13    16:00:00    REGULAR 4087738.0   1406741.0   04-23-13    20:00:00    REGULAR 4088682.0   1406813.0
3   A002    R051    02-00-00    04-24-13    00:00:00    REGULAR 4088879 1406839 04-24-13    04:00:00    REGULAR 4088890.0   1406845.0   04-24-13    08:00:00    REGULAR 4088934.0   1406951.0   04-24-13    12:00:00    REGULAR 4089105.0   1407209.0   04-24-13    16:00:00    REGULAR 4089378.0   1407269.0   04-24-13    20:00:00    REGULAR 4090319.0   1407336.0   04-25-13    00:00:00    REGULAR 4090535.0   1407365.0   04-25-13    04:00:00    REGULAR 4090550.0   1407370.0
4   A002    R051    02-00-00    04-25-13    08:00:00    REGULAR 4090589 1407469 04-25-13    08:57:03    DOOR OPEN   4090629.0   1407591.0   04-25-13    08:58:01    LOGON   4090629.0   1407591.0   04-25-13    09:01:08    LGF-MAN 4090629.0   1407591.0   04-25-13    09:01:53    LOGON   4090629.0   1407591.0   04-25-13    09:02:02    DOOR CLOSE  4090629.0   1407591.0   04-25-13    09:02:04    DOOR OPEN   4090629.0   1407591.0   04-25-13    09:02:31    DOOR CLOSE  4090629.0   1407591.0
5   A002    R051    02-00-00    04-25-13    09:02:32    DOOR OPEN   4090629 1407591 04-25-13    09:07:21    LOGON   4090629.0   1407591.0   04-25-13    09:12:12    LGF-MAN 4090642.0   1407592.0   04-25-13    09:12:20    DOOR CLOSE  4090642.0   1407592.0   04-25-13    12:00:00    REGULAR 4090743.0   1407723.0   04-25-13    16:00:00    REGULAR 4091064.0   1407793.0   04-25-13    20:00:00    REGULAR 4092044.0   1407840.0   04-26-13    00:00:00    REGULAR 4092314.0   1407859.0
6   A002    R051    02-00-00    04-26-13    04:00:00    REGULAR 4092325 1407861 04-26-13    08:00:00    REGULAR 4092363.0   1407958.0   04-26-13    12:00:00    REGULAR 4092541.0   1408225.0   04-26-13    16:00:00    REGULAR 4092837.0   1408285.0   04-26-13    20:00:00    REGULAR 4093823.0   1408341.0   None    None    None    NaN NaN None    None    None    NaN NaN None    None    None    NaN NaN

我当前的低效函数如下所示：

def cleanData(dataFrame):

    tempDf = dataFrame

    tempColName = ['date','time','description','entries','exit','c/a','unit', 'scp']
    finalColName = ['c/a','unit','scp','date','time','description','entries','exit']

    tempDf1 = tempDf.iloc[:,:8]
    tempDf1.dropna(inplace=True)
    tempDf1.columns = finalColName

    tempDf2 = tempDf.iloc[:,8:13]
    tempDf2['c/a'] = tempDf['C/A']
    tempDf2['unit'] = tempDf['UNIT']
    tempDf2['scp'] = tempDf['SCP']
    tempDf2.dropna(inplace=True)
    tempDf2.columns = tempColName

    tempDf3 = tempDf.iloc[:,13:18]
    tempDf3['c/a'] = tempDf['C/A']
    tempDf3['unit'] = tempDf['UNIT']
    tempDf3['scp'] = tempDf['SCP']
    tempDf3.dropna(inplace=True)
    tempDf3.columns = tempColName

    tempDf4 = tempDf.iloc[:,18:23]
    tempDf4['c/a'] = tempDf['C/A']
    tempDf4['unit'] = tempDf['UNIT']
    tempDf4['scp'] = tempDf['SCP']
    tempDf4.dropna(inplace=True)
    tempDf4.columns = tempColName

    tempDf5 = tempDf.iloc[:,23:28]
    tempDf5['c/a'] = tempDf['C/A']
    tempDf5['unit'] = tempDf['UNIT']
    tempDf5['scp'] = tempDf['SCP']
    tempDf5.dropna(inplace=True)
    tempDf5.columns = tempColName

    tempDf6 = tempDf.iloc[:,28:33]
    tempDf6['c/a'] = tempDf['C/A']
    tempDf6['unit'] = tempDf['UNIT']
    tempDf6['scp'] = tempDf['SCP']
    tempDf6.dropna(inplace=True)
    tempDf6.columns = tempColName

    tempDf7 = tempDf.iloc[:,33:38]
    tempDf7['c/a'] = tempDf['C/A']
    tempDf7['unit'] = tempDf['UNIT']
    tempDf7['scp'] = tempDf['SCP']
    tempDf7.dropna(inplace=True)
    tempDf7.columns = tempColName

    tempDf8 = tempDf.iloc[:,38:43]
    tempDf8['c/a'] = tempDf['C/A']
    tempDf8['unit'] = tempDf['UNIT']
    tempDf8['scp'] = tempDf['SCP']
    tempDf8.dropna(inplace=True)
    tempDf8.columns = tempColName

    placeHolderDf = pd.concat([tempDf2,tempDf3,tempDf4,tempDf5,tempDf6,tempDf7,tempDf8])
    placeHolderDf = placeHolderDf[['c/a','unit','scp','date','time','description','entries','exit']]
    fullData = pd.concat([tempDf1,placeHolderDf])
    fullData['date'] = pd.to_datetime(fullData['date'])

    return fullData.reset_index(drop=True)

具有正确的最终输出，如：

    c/a     unit    scp        date         time    description entries exit
0   A002    R051    02-00-00    2013-04-20  00:00:00    REGULAR 4084276 1405308
1   A002    R051    02-00-00    2013-04-21  08:00:00    REGULAR 4085336 1405603
2   A002    R051    02-00-00    2013-04-22  16:00:00    REGULAR 4086116 1406242
3   A002    R051    02-00-00    2013-04-24  00:00:00    REGULAR 4088879 1406839
4   A002    R051    02-00-00    2013-04-25  08:00:00    REGULAR 4090589 1407469

非常感谢您的帮助


Tags： columnsnonetrueunitscpregularinplaceiloc
2条回答
网友
1楼 · 编辑于 2024-10-06 16:27:40

希望这符合你的要求

pd.wide_to_long将自动对列名进行分组

list_melt =  ['DATE','TIME','DESC','ENTRIES','EXITS']
list_id = ['C/A','UNIT','SCP']
# pd.wide_to_long id has to be unique
df['count'] = df.groupby(list_id)[list_id].cumcount()
df_melt = pd.wide_to_long(df, stubnames = list_melt, i = ['C/A','UNIT','SCP','count'], j = 'value').reset_index()
df_melt.drop(columns = ['count', 'value'], axis = 1, inplace = True)
df_melt.columns = ['c/a','unit','scp','date','time','description','entries','exit']

这是一个代理输出：

list_cols = ['C/A','UNIT','SCP','DATE1','TIME1','DESC1','ENTRIES1','EXITS1','DATE2','TIME2','DESC2',\
 'ENTRIES2','EXITS2','DATE3','TIME3','DESC3','ENTRIES3','EXITS3']

df = pd.DataFrame([[1] * 8 + [1,2,3,4,5] + [6,7,8,9,10]], columns = list_cols)

df_melt
  c/a   unit    scp date    time description entries exit
0   1     1       1    1    1             1    1      1
1   1     1       1    1    2             3    4      5
2   1     1       1    6    7             8    9      10
网友
2楼 · 编辑于 2024-10-06 16:27:40

您可以尝试：

import io
import pandas as pd

s="""C/A,UNIT,SCP,DATE1,TIME1,DESC1,ENTRIES1,EXITS1,DATE2,TIME2,DESC2,ENTRIES2,EXITS2
   A002,R051,02-00-00,04-20-13,00:00:00,REGULAR,4084276,1405308,04-20-13,04:00:00,REGULAR,4084308.0,1405312.0
   A002,R051,02-00-00,04-25-13,09:02:32,DOOR OPEN,4090629,1407591,04-25-13,09:07:21,LOGON,4090629.0,1407591.0
   A002,R051,02-00-00,04-26-13,04:00:00,REGULAR,4092325,1407861,04-26-13,08:00:00,REGULAR,4092363.0,1407958.0
"""

df = pd.read_csv(io.StringIO(s), sep=',')

col_names = ['C/A', 'UNIT', 'SCP', 'DATE', 'TIME', 'DESC', 'ENTRIES', 'EXITS']

i = 0
nr = 2 # change to 8 with your file
df_dict = dict()
while i < nr:
    i+=1
    df_dict[i] = df.loc[:, [column for column in df.columns[:3]] + [column for column in df.columns if column.endswith(str(i))]]
    new_cols = {x: y for x, y in zip(df_dict[i], df_new.columns)}
    df_dict[i] = df_dict[i].rename(columns=new_cols)

df_new = pd.concat(df_dict.values())
print(df_new)

相关问题 更多 >

    编程相关推荐