ValueError：无法将操作数与形状一起广播到矢量化函数

Question

我正在 Google Colab 中制作一个 python 程序，在这个 Colab 中，我想创建一个函数，将两个 Numpy 数组、一个 panda Dataframe 和一个字符串发送到一个 Numpy 矢量化函数。当我执行这段代码时，我收到以下错误：ValueError: operands could not be broadcast together with shapes (6501398,) (6501398,) (462650,11) ()。

这是我使用的代码：

def calc_gem_bouwjaar(postcode_van_ascii, postcode_tot_ascii, BAG_grouped, foo_string):
    if (postcode_van_ascii != postcode_tot_ascii):
    total = 0
    amount = 0
    postcode_van_ascii = add_to_barcode(val=postcode_van_ascii, pos=1) # this just increments the value by one, and handles overflows in postal code format.

    while postcode_van_ascii != postcode_tot_ascii:
      try:
        # at is de snelste manier om data uit de dataframe te halen https://github.com/pandas-dev/pandas/issues/6683#issuecomment-38305770. 
        # df.ix lijkt nog sneller maar is sinds 0.23.4 deprecated https://pandas.pydata.org/pandas-docs/version/0.23.4/generated/pandas.DataFrame.ix.html
        total += round(BAG_grouped.at[postcode_van_ascii, foo_string])
        amount += 1
      except:
        pass

      postcode_van_ascii = add_to_barcode(val=postcode_van_ascii, pos=1) # this just increments the value by one, and handles overflows in postal code format.

    try:
      return round(total/amount)
    except:
      # Error in division, probably no excisting data in grouped dataframe
      return 0

  else:
    try:
      return int(round(BAG_grouped.at[postcode_van_ascii, foo_string]))
    except:
      return 0


def foo(foo_string):
    # do other stuff
    BAG_grouped = BAG_df.groupby('postcode_ascii').mean()
    vconvert = np.vectorize(calc_gem_bouwjaar)

    postcode_van_array = np.asarray(Netbeheerders_df["POSTCODE_VAN_ASCII"]).astype(int)
    postcode_tot_array = np.asarray(Netbeheerders_df["POSTCODE_TOT_ASCII"]).astype(int)

    # foo_string = 'bouwjaar'
    result = vconvert(postcode_van_array, postcode_tot_array, BAG_grouped, foo_string).astype(int)  #<--- ValueError: operands could not be broadcast together with shapes (6501398,) (6501398,) (462650,11) ()


foo('bouwjaar')

为什么我的代码会产生这个错误，我该如何解决这个问题？

Answer 1

您的函数calc_gem_bouwjaar 通过广播对您的向量进行操作。由于您的矢量形状不可广播（特别是 (6501398,) (462650,11) 和 ()），因此会引发错误。您必须查看您的功能，看看您在哪里尝试进行需要广播的操作并修复相应的形状/操作。没有函数的内容，我们可以添加的东西不多。