将熊猫数据集转换为numpy

Question

我正在尝试读取我的数据集并将其转换为numpy。我发现此错误如何解决。我试图更改数据的属性，但存在相同的问题

def df_to_npy(df, left_context=5, right_context=5):
    print 'Converting DataFrame to Numpy...'
    cntlab = 0
    cntfeat = 0
    for idx, row in df.iterrows():
        cntlab += row['labs'].shape[0]
        cntfeat += row['feats'].shape[0]

    if cntfeat == cntlab:
        cnt = cntfeat

AttributeError：“浮动”对象没有属性“ shape”