【Maxcompute】实现根据gps数据计算停留点udaf函数

1.梳理、总结经纬度处理在Maxcompute平台上的实战应用，如根据gps轨迹数据计算停留点信息。
2.欢迎批评指正，跪谢一键三连！

1.参考代码

• #!/usr/bin/env python
  # coding: utf-8
  
  import numpy as np
  from odps.udf import annotate
  from odps.udf import BaseUDAF
  from datetime import datetime
  
  @annotate('string,string,double,string->string')
  class ret_tld_by_gps(BaseUDAF):
      """计算停留点"""
      def new_buffer(self):
          return []
      def iterate(self, buffer, lati, longi, speed, gpstime):
          buffer.append([lati, longi, speed, gpstime])
      def merge(self, buffer, pbuffer):
          buffer += pbuffer
      def terminate(self, buffer): 
          buf = np.array(buffer)
          # try:
          buf2 = buf[:, 3]
          # return '%s %s %d %d' % (type(buf), type(buffer), buf.shape[0], buf.shape[1])
          index = np.lexsort((buf2,))
          buf3 = buf[index]
          # return '%s %d %d' % (','.join(buf3[0]), buf3.shape[0], buf3.shape[1])
          tld = []
          cnt, total_cnt, _flag = 0, 0, 0
          lati ,longi ,speed ,first_gpstime ,last_gpstime  = '','',0.0,'',''
          for rw in buf3:
              total_cnt += 1
              if float(rw[2])>0:
                  if _flag==1: # 速度变化点
                      if cnt==1:
                          last_gpstime = first_gpstime
                      # return '114, %d %s %d ' % (_flag, str(rw[2]), _flag)
                      # if (first_gpstime<>'') and (last_gpstime<>''):
                      diff_time = (datetime.strptime(last_gpstime, "%Y-%m-%d %H:%M:%S")-datetime.strptime(first_gpstime, "%Y-%m-%d %H:%M:%S")).seconds
                      # tld.append(','.join([lati, longi, first_gpstime, last_gpstime, '%.2f' % (diff_time/60.0) , str(cnt)]))
                      if (diff_time>=600) and (diff_time<=2400):
                          tld.append(','.join([lati, longi, first_gpstime, last_gpstime, '%.2f' % (diff_time/60.0) , str(cnt)]))
                      _flag, cnt = 0, 0
              elif float(rw[2])<=0: # 起始停留时间
                  cnt += 1
                  if _flag==0: # 持续为0点轨迹点
                      lati, longi = rw[0], rw[1]
                      first_gpstime, _flag = rw[3], 1
                  else:
                      last_gpstime = rw[3]
          if cnt!=1:
              # print('')
              # if (first_gpstime<>'') and (last_gpstime<>''):
              diff_time = (datetime.strptime(last_gpstime, "%Y-%m-%d %H:%M:%S")-datetime.strptime(first_gpstime, "%Y-%m-%d %H:%M:%S")).seconds
              if (diff_time>=600) and (diff_time<=2400):
                  tld.append(','.join([lati, longi, first_gpstime, last_gpstime, '%.2f' % (diff_time/60.0) , str(cnt)]))
          # return '%s' % (str(total_cnt))
          return '%s' % (';'.join(tld))