python3__零散问题处理

àì夳堔傛蜴生んèń 2023-02-16 10:54 52阅读 0赞

**目录**

1.dataframe中可迭代元素因长度不同修改时报错的问题

2.numpy数组中nan元素的替换

--------------------

# 1.dataframe中可迭代元素因长度不同修改时报错的问题 #

> **原始代码**：将商品标题对应特征词列表转换为向量并重新赋值

for i in range(itemsData.shape[0]):
            words = itemsData["name"][i].split(" ")
            tmp1 = np.zeros(w2c.wv.vector_size)
            for word in words:
                tmp1 += w2c.wv.get_vector(word)
            itemsData["name"][i] = tmp1/len(words)
        itemsData.to_csv("word2vec_vector/itemsData_after_vector.csv", sep=",", encoding="utf-8", index=False, header=True)
    
    
    # 报错
    # pandas ValueError: Must have equal len keys and value when setting with an iterable

> **修改方式**：将待赋值的可迭代对象看成一个整体并置换

for i in range(itemsData.shape[0]):
            words = itemsData["name"][i].split(" ")
            tmp1 = np.zeros(w2c.wv.vector_size)
            for word in words:
                tmp1 += w2c.wv.get_vector(word)
            itemsData.set_value(i, "name", tmp1/len(words))
        itemsData.to_csv("word2vec_vector/itemsData_after_vector.csv", sep=",", encoding="utf-8", index=False, header=True)

# 2.numpy数组中nan元素的替换 #

# ==================================================================nan元素的替换
    from numba import jit
    import numpy as np
    import time
    
    
    @jit
    def nan_to_statistic_value(arr, method="mean"):
        """将数组arr中的nan替换成统计量"""
        nan_num = np.count_nonzero(np.isnan(arr))
        if nan_num == 0:
            return arr
        else:
            for i in range(arr.shape[1]):
                tmp_col = arr[:, i]
                if np.count_nonzero(np.isnan(tmp_col)) != 0:
                    tmp_not_nan_col = tmp_col[tmp_col == tmp_col]
                    if method == "mean":
                        tmp_col[np.isnan(tmp_col)] = tmp_not_nan_col.mean()
                    elif method == "other":
                        # TODO 可以继续丰富
                        return arr
    
            return arr
    
    
    def main():
        start = time.time()
        arr = np.arange(24).reshape((4, 6)).astype(float)
        arr[2, 3:] = np.nan
        print("arr:\n", arr)
        print("new_arr:\n", nan_to_statistic_value(arr))
        print("time cost:", time.time() - start)
    
    
    if __name__ == "__main__":
        main()