如何使用Python将csv文件中的datetime对象拆分为30分钟数组，并将每个数组导出为新的csv-fi

Question

我有一个10赫兹的.csv数据文件，它的日期/时间戳在同一列中，格式如下：2014-07-05 12:01:00.0到2014-07-06 12:00:59.9，作为文件中一天的数据。我需要在24小时内（每天48小时）分成30分钟的时间段，12:01:00.0-12:29:59.9，12:30:00.0-12:59:59.9，以此类推。每个拆分的30分钟块需要导出为自己的文本文件。我对python很陌生（第3天），我正在尽我最大的努力，但是我在旋转我的轮子。我在实习，真的需要帮助。我不是程序员，只是一个试图用python来完成这项任务的化学家。我试图按行进行拆分（rows=18000，30分钟），但由于我的数据在12:00:01开始时并不均匀，这会扰乱我的行拆分计算，并且无法提供精确的30分钟分割。我被告知需要一个日期时间对象来转换成字符串。任何指导或帮助将不胜感激。提前谢谢。下面是我已经开始为日期时间对象修改它的内容，但确实需要一些指导：

import csv
import re
import os
import datetime
import numpy as np
filename = 'C:\Users\Jason\Documents\Flux Data Files\HL14_175.csv'
f = open('C:\Users\Jason\Documents\Flux Data Files\Output Flux Split 30 mins Data    Files\HL14_175_split0.csv','wb')
writer = csv.writer(f,delimiter = ',')
with open(filename,"r") as datafile:
    r = csv.reader(datafile,delimiter = ",")
    timestamp = datetime.datetime.strptime("2014-07-05", "%Y-%m-%d %H:%M:%S:%f")
    recordnumber = []
    sonic1 = []
    sonic2 = []
    sonic3 = []
    temperature = []
    for row in r:
        timestamp.append((row[0]))
        recordnumber.append(float(row[1]))
        sonic1.append(float(row[2]))
        sonic2.append(float(row[3]))
        sonic3.append(float(row[4]))
        temperature.append(float(row[5]))
    timestamp = np.array(timestamp)
    recordnumber = np.array(recordnumber)
    sonic1 = np.array(sonic1)
    sonic2 = np.array(sonic2)
    sonic3 = np.array(sonic3)
    temperature = np.array(temperature)
datetime.strptime(date_string, format) 
#row_count = 863998
row_count = sum(1 for row in csv.reader(open(filename)))
lines = row_count/18001.0
timestamp_split = np.array_split(timestamp,lines)
recordnumber_split = np.array_split(recordnumber,lines)
sonic1_split = np.array_split(sonic1,lines)
sonic2_split = np.array_split(sonic2,lines)
sonic3_split = np.array_split(sonic3,lines)
temperature_split = np.array_split(temperature,lines)
dataout =  np.column_stack((timestamp_split[0],recordnumber_split[0],sonic1_split[0],sonic2_split[0],sonic3_split[0],temperature_split[0]))
writer.writerows(dataout)
f.close()
print('Flux Data Split Complete')

下面是一个示例数据文件：

^{pr2}$

Answer 1

这不是完整的解决方案，因为正确的日期转换仍然存在问题。在

我使用文本数据来模拟从csv读取。在

我对pandas是个初学者，所以也许有人能把它做得更好。在

import pandas as pd
import StringIO

data = '''6/24/2014 0:01,3583014,-59,-62,-9,296.51
01:00.1,3583015,-69,-68,16,296.54
01:00.2,3583016,-62,-59,36,296.56
01:00.3,3583017,-77,-45,26,296.56
01:00.4,3583018,-47,-50,36,296.56
01:00.5,3583019,-48,-70,27,296.51
01:00.6,3583020,-71,-60,28,296.54
01:00.7,3583021,-69,-73,24,296.52
01:00.8,3583022,-61,-69,15,296.49
01:00.9,3583023,-56,-68,8,296.52
6/24/2014 0:01,3583024,-65,-42,-5,296.56
01:01.1,3583025,-71,-33,-11,296.56
6/24/2014 0:31,3583014,-59,-62,-9,296.51
31:00.1,3583015,-69,-68,16,296.54
31:00.2,3583016,-62,-59,36,296.56
31:00.3,3583017,-77,-45,26,296.56
31:00.4,3583018,-47,-50,36,296.56
31:00.5,3583019,-48,-70,27,296.51
31:00.6,3583020,-71,-60,28,296.54
31:00.7,3583021,-69,-73,24,296.52
31:00.8,3583022,-61,-69,15,296.49
31:00.9,3583023,-56,-68,8,296.52
6/24/2014 0:31,3583024,-65,-42,-5,296.56
31:01.1,3583025,-71,-33,-11,296.56'''

# reading from CSV

df = pd.DataFrame.from_csv(StringIO.StringIO(data), index_col=None, header=None)

#print df

# converting "wierd" date format - still can be problem

date = None
minut = None
second = 59

def change_date(line):
    global date, minut, second
    a = line.split(':')
    if len(a[0]) > 2:
        if a[0] != date or a[1] != minut:
           second = 59
        date = a[0]
        minut = a[1]
        second = (second + 1) % 60
        return "%s:%02d.0" % (line, second)
        #return line
    else:
        return date + ":" + line

df[0] = df[0].map(change_date)

#print df
#print df.dtypes

# converting string with date and time to object datetime

df[0] = pd.DatetimeIndex(df[0])

#print df.dtypes

# groub by date (year,month,day,hour) and minute (minute<30)

g = df.groupby( df[0].map(lambda t:(t.strftime("%Y_%m_%d_%H_") + ("00" if t.minute<30 else "30") )) )

# print groups

for name, group in g:
    print 'name:', name
    print group
    group.to_csv(name + ".csv") # write groups to files

结果

^{pr2}$