Question

我有一个包含CSV数据的Pandas DataFrame，如下所示：

time animal_id x y

0 1 120 205.29 217.76

1 1 100 169.99 128.78

2 1 200 590.33 505.89

3 1 300 645.15 611.94

4 1 350 866.06 851.76

文件的尺寸为216005行和4个属性/列。

我已经根据“ animal_id”对文件进行了分组，之后又添加了一些我必须计算的其他属性（即距离，average_speed，average_acceleration，positive_acceleration，方向）。

此代码为：

# Apply grouping using 'animal_id' attribute-
data_animal_id = data.groupby('animal_id')

# A dictionary object to hold all groups obtained using group by-
data_animal_id_groups = {}


# Get each animal_id's data from grouping performed-
for animal_id in data_animal_id.groups.keys():
    data_animal_id_groups[animal_id] = data_animal_id.get_group(animal_id)

# To reset index for each group-
for animal_id in data_animal_id_groups.keys():
    data_animal_id_groups[animal_id].reset_index(drop=True, inplace=True)


# Add additional attributes/columns to each groups-
for aid in data_animal_id_groups.keys():
    data = [0 for x in range(data_animal_id_groups[aid].shape[0])]

    data_animal_id_groups[aid] = data_animal_id_groups[aid].assign(distance=data)
    data_animal_id_groups[aid] = data_animal_id_groups[aid].assign(average_speed=data)
    data_animal_id_groups[aid] = data_animal_id_groups[aid].assign(average_acceleration=data)
    data_animal_id_groups[aid] = data_animal_id_groups[aid].assign(positive_acceleration=data)
    data_animal_id_groups[aid] = data_animal_id_groups[aid].assign(direction=data)

我现在正在尝试使用以下逻辑来计算方向：

for aid in data_animal_id_groups.keys():
    print("\nComputing Distance & Direction for Animal ID = {0}\n".format(aid))
    for i in range(1, data_animal_id_groups[aid].shape[0] - 1):     

        x1 = data_animal_id_groups[aid].iloc[i, 2]
        y1 = data_animal_id_groups[aid].iloc[i, 3]
        x2 = data_animal_id_groups[aid].iloc[i + 1, 2]
        y2 = data_animal_id_groups[aid].iloc[i + 1, 3]

        # Compute distance between 2 points-
        distance = math.sqrt(math.pow((x2 - x1), 2) + math.pow((y2 - y1), 2))

        # Compute the direction in DEGREES-
        direction = math.degrees(math.atan((y2 - y1) / (x2 - x1)))

        if math.isnan(direction):
            data_animal_id_groups[aid].loc[i, 'Direction'] = 0
            # animal_id.loc[i, 'Direction'] = 0
        else:
            data_animal_id_groups[aid].loc[i, 'Direction'] = direction

但是，这需要很长时间才能计算出来。因此，我尝试使用以下numpy代码：

direction = np.rad2deg(np.arctan((data_animal_id_groups[312]['y'] - data_animal_id_groups[312]['y'].shift(periods = 1, axis = 0)), (data_animal_id_groups[312]['x'] - data_animal_id_groups[312]['x'].shift(periods = 1, axis = 0))))

但是，这给出了错误：

RecursionError：调用a时超出了最大递归深度 Python对象

我该如何解决？

谢谢

熊猫列计算-递归错误

0 个答案: