问当分钟重复时，从数据帧到时间序列

EN

Time (min)   MeanWait (min, single default patient)

...          ...
420          5.2
421          4.9
422          4.3
423          4.2
...
...
820          11.39
821          11.41
822          11.41
823          11.09
824          10.7
825          10.69
...          ...

import datetime

# The day timescale is from 0 to 1440 minutes and then resets for day 2.
# The input-textfile can have 24h (0-1440) or continues (e.g. 0-4320 == 3 days) timescaling for x-axis.

# Testset for dataprocessing (day 1 and day2 data)
datas = ['Time(minutes)       RemainingTime(minutes)', 
        '420              :                      5',
        '420              :                      10',
        '420              :                      8',
        '421              :                      4',
        '421              :                      9',
        '421              :                      7',
        '830              :                      8',
        '830              :                      4',
        '340              :                      3',
        '340              :                      5',
        '340              :                      4',
        '351              :                      10',
        '351              :                      7',
        '420              :                      9',
        '420              :                      7',]

def sort_data(scr):

    raw_data            = {}
    day_minute_counter  = 0
    current_list        = []
    day_in_minutes      = (24 * 60)
    elapsed_days_min    = 0           # during processing this holds value in minutes
    processed_days      = 1
    data_from_exception = {}
    count_exceptions    = 0

    for row in scr:
        print row

        try:
            # the following steps take ito account that lapsed time is linear for a single day.
            # each row is being searched for ":" which identifies teh row as having integers or floats.
            x_value, y_value = row.split(":")

#            print 'xy_values : %s, %s' % (x_value, y_value)

            # clipping trailing whitespaces from both ends.
            x_value = x_value.strip(' ')
            y_value = y_value.strip(' ')

            # string > integer conversion
            x_val = int(x_value)
            y_val = int(y_value)

            # set each x-axis timepoint only once.
            if day_minute_counter == 0:
                print 'Start', day_minute_counter, x_val

                day_minute_counter = x_val

            # zipping: append all y-axis datapoint that belong to single x-axispoint
            if day_minute_counter == x_val:
                print 'Append', day_minute_counter, x_val
                current_list.append(y_val)

            # add x,y-axis data to the datalist
            if day_minute_counter < x_val:

                print 'Done', day_minute_counter, x_val, current_list

                raw_data[(day_minute_counter + elapsed_days_min)] = current_list

                day_minute_counter = x_val

                # new list for the next point in the "day_minute_counter".
                current_list      = []  
                current_list.append(y_val)

            # correct x-axis "next-day" time difference.
            if day_minute_counter > x_val:

                processed_days += 1

                print 'Next Day Marker', day_minute_counter, x_val, current_list

                raw_data[(day_minute_counter+ elapsed_days_min)] = current_list

                elapsed_days_min += day_in_minutes
                # reset day_minute_counter because a day has elapsed.
                day_minute_counter = 0
                print 'elapsed_day in minutes : ', elapsed_days_min

        except ValueError:
            #get axis information
            count_exceptions += 1
            data_from_exception[count_exceptions] = row
#            print 'Graph info or "none integer" information collected:\n\n%s > %s\n' % (count_exceptions, row)

    # End of datablock : add the last x,y datapoints without known what EOF marker is being used.
    raw_data[day_minute_counter] = current_list

    print '\nRaw Data   : %s\nOther info : %s\n ' % (raw_data, data_from_exception)

    return (raw_data, processed_days, data_from_exception)

def calc_mean(scr):

    days = scr[1]
    minutes = (days * 24 * 60)
    missing_datapoints = []
    result = []
    print 'Dataset spans a total of "%s" minutes.\n' % minutes

    data = scr[0]

    for x_datapoint in range(1, minutes):

        meanwait  = 0.0
        totalwait = 0

        try:
            # process data from sorte_data.

#            print 'datapoint', x_datapoint  # shows only the absent datapoints on x-axis.
            dataset   = data[x_datapoint]
#            print 'datapoint', x_datapoint  # shows only the available datapoints on x-axis.

            total_values = len(dataset)
            for value in dataset:
                totalwait += value

            meanwait = float(totalwait) / float(total_values)

            x = x_datapoint
            y = meanwait

            result.append((x, y))

            print 'Patient meanwaiting time per Timepoint %s : %.03f' % (x_datapoint, meanwait)

        except Exception:
            missing_datapoints.append(x_datapoint)
#            print 'Patient meainwaiting time "%s" is not available.' % x_datapoint

    return result

def main():

    # open file code here and use readlines to import data to "datas"
    #
    # datas = ...

    ct = str(datetime.datetime.now())[0:23]

    print '%s --> Collecting patient waittime data from Time Series.\n' % ct

    sorted_data = sort_data(datas)   # used template date from this script.

    print 'Processing data to obtain main values'

    the_result = calc_mean(sorted_data)

    print '\nProcessing Finished. Here is the result :\n\n%s' % the_result

    # create new file and store result or keep processing to PDF in matplotlib

if __name__ == '__main__':

    main()