data_array = np.loadtxt(filename,   #文件名
              delimiter=',', #分隔符
              skiprows=1,  #跳过第一行
              dtype=bytes,   #数据类型
              usecols=use_col_index_lst).astype(str) #用指定列
  '''
  I think np.loadtxt("tile", dtype=bytes, delimiter="\n").astype(str) might work, 
  but I agree completely with the overkill point (解决python2,python3编码不一样的问题)