陷入机器学习应用的数据预处理代码

Question

我目前正在开发一个机器学习应用程序。请在此代码中帮助我 - 当我上传大数据集时，我遇到了一个错误。

代码如下:

prep_file = Prepross.objects.get_or_create(
                filename = CurrentFile.objects.order_by('-id')[0].filename,
                coltype = request.POST.getlist('coltype'),
                assvar = request.POST.getlist('assvar'),
                missingvalues = request.POST.getlist('missingvalues'),
                trainingset_size = request.POST['trainingset_size'],
                featscaling = request.POST.getlist('featscaling'))

然后：

        # Get dataframe and change data type
        context = {}
        file_name = CurrentFile.objects.order_by('-id')[0].filename
        coltype = request.POST.getlist('coltype')
        coltype = dict([i.split(':', 1) for i in coltype])
        df = pd.read_csv(os.path.join('media\downloaded', file_name), dtype= coltype)
        row_count = df.count()[0]

        # Keep only selected columns
        assvar = request.POST.getlist('assvar')
        xcols0 = [s for s in assvar if ":X" in s]
        xcols = [i.split(':', 1)[0] for i in xcols0]
        ycol0 = [s for s in assvar if ":y" in s]
        ycol = [i.split(':', 1)[0] for i in ycol0]
        cols = xcols + ycol
        df = df[cols]

        xcols = ', '.join(xcols)
        ycol = ', '.join(ycol)
        missing = request.POST.getlist('missingvalues')
        missing = ', '.join(missing)
        trainingset_s = request.POST.getlist('trainingset_size')
        trainingset_s = ', '.join(trainingset_s)
        testset_s = 100 - int(trainingset_s)
        feat =  request.POST['featscaling']

        # Taking care of missing data
        if missing == "no":
            if len(df) != len(df.dropna()):
                context['selecty'] = 'Your data seem to have Missing Values'
            else:
                df = df.dropna()

        # Return error if columns are not selected
        if len(ycol0) != 1:
            context['selecty'] = 'Please select one y variable'

        elif len(xcols0) < 1:
            context['selecty'] = 'Please select one or more X variables'

错误是：

 File "C:\Users\Admin\PycharmProject\mlapp\views.py", line 81, in post
    coltype = dict([i.split(':', 1) for i in coltype])
ValueError: dictionary update sequence element #0 has length 1; 2 is required
[20/Aug/2019 16:43:39] "POST /preprocessing/ HTTP/1.1" 500 81482

带有关于错误的附加行和信息：

/预处理/字典更新序列元素#0处的ValueError的长度为1；2是必需的请求方法:POST请求URL:http://127 . 0 . 0 . 1:8000/preprocessing/Django版本:2.2.4异常类型:ValueError异常值:< br >字典更新序列元素#0的长度为1；2是必需的异常位置:C:\ Users \ Admin \ pycharm project \ ml app \ views . py in post，第81行Python可执行文件:C:\ Users \ Admin \ AppData \ Local \ Programs \ Python \ Python 37 \ Python . exe Python版本:3.7.3 Python路径:< br >[' C:\ Users \ Admin \ pycharm project \ freed '，' C:\ Users \ Admin \ AppData \ Local \ Programs \ Python 37 \ Python 37 . zip '，' C:\ Users \ Admin \ AppData \ Local \ Programs \ Python \ Python \ Programs \ Python \ Python

Answer 1

你好像想编一本字典

 coltype = dict([i.split(':', 1) for i in coltype])

但要制作字典，需要两个参数Key和Value。你只是传递了一个参数。也许，这就是为什么，它显示了这个错误。

Answer 2

错误位于此行：

coltype = dict([i.split(':', 1) for i in coltype])

并且表示其中一个元素（可能是第一个）的长度为1。这意味着没有“：”可以拆分。你期望每个元素都是“a： b”的形式，但显然其中一个不是这种情况。

在假定数据具有特定格式之前，应始终对其进行验证。在您的情况下，是请求中的元素之一。POST.getlist（“coltype”）没有您期望的格式。