Question

有以下型号：

class Parameter (models.Model):
    id_parameter = models.IntegerField(primary_key=True)
    par_rollennr = models.IntegerField(default=0)
    par_definition_id = models.IntegerField(default=0) #not FK
    par_name = models.CharField(max_length=200)


class Measurements (models.Model):
    id_measurement = models.AutoField(primary_key=True)
    par_value = models.IntegerField(default=0)
    line = models.ForeignKey(Line, on_delete=models.CASCADE, null=True)
    order = models.ForeignKey(Order, on_delete=models.CASCADE, null=True)
    recipe = models.ForeignKey(Recipe, on_delete=models.CASCADE, null=True)
    parameter = models.ForeignKey(Parameter, on_delete=models.CASCADE, null=True)

我写下如下： def handle_parameters_upload（请求，文件）：

wb = openpyxl.load_workbook(file, read_only=True)
first_sheet = wb.get_sheet_names()[0]
ws = wb.get_sheet_by_name(first_sheet)

recipe, created = Recipe.objects.get_or_create(par_recipe=ws["B2"].value)

line, created = Line.objects.get_or_create(par_machine=ws["C2"].value)

order, created = Order.objects.get_or_create(par_fa=ws["D2"].value)

data = []
data_par = []
_id = 1
for row in ws.iter_rows(row_offset=1):
    parameter = Parameter()
    parameter.id_parameter = _id
    _id += 1
    parameter.par_rollennr = row[5].value
    parameter.par_definition_id = row[6].value
    parameter.par_name = row[7].value
    data_par.append(parameter)

    measurements = Measurements()
    measurements.par_value = row[8].value
    measurements.line = line
    measurements.order = order
    measurements.parameter = parameter
    measurements.recipe = recipe

    data.append(measurements)
# Bulk create data
Measurements.objects.all().delete()
Parameter.objects.all().delete()
Parameter.objects.bulk_create(data_par)
Measurements.objects.bulk_create(data)
return True

如何避免参数表中的记录重复并且不丢失ID的依赖关系。参数是文件中的3个字段，每一行都有其自己的字段，但总数不超过1052，并且每1052个条目分别重复一次。看起来像这样：

rollennr | definitionid | name | value

其中rollennr，definitionid，name是模型参数字段。以后，我希望通过以下查询来还原此信息：

select * from Measurements as m
join Parameter as p on m.parameter_id = p.id

该问题的一种临时解决方案，该方法在很长一段时间内击败了数据：

def handle_parameters_upload_v2(request, file):

    wb = openpyxl.load_workbook(file, read_only=True)
    first_sheet = wb.get_sheet_names()[0]
    ws = wb.get_sheet_by_name(first_sheet)

    recipe, created = Recipe.objects.get_or_create(par_recipe=ws["B2"].value)

    line, created = Line.objects.get_or_create(par_machine=ws["C2"].value)

    order, created = Order.objects.get_or_create(par_fa=ws["D2"].value)

    Measurements.objects.all().delete()
    Parameter.objects.all().delete()

    data = []
    data_par = []

    _id = 0
    for row in ws.iter_rows(row_offset=1):
        _id += 1
        parameter = dict(par_rollennr=row[5].value, par_definition_id=row[6].value, par_name=row[7].value)
        if parameter not in data_par:
            # data_par_all.append(parameter)
            data_par.append(parameter)
        measurements = dict(par_value=row[8].value, line=line, order=order, parameter=parameter, recipe=recipe)
        data.append(measurements)
    else:
        write_param(data_par, data)
    return True


def write_param(data_par, data):
    _id = 0
    for i in data_par:
        i['id_parameter'] = _id
        _id += 1
    Parameter.objects.all().delete()
    parameter = [Parameter(**i) for i in data_par]
    created_param = Parameter.objects.bulk_create(parameter)
    # data_par_obj_all = data_par_obj_all + created_param
    for obj in data:
        obj['parameter'] = find_parameter(obj, created_param)
    Measurements.objects.all().delete()
    measurements = [Measurements(**i) for i in data]
    i = 0
    while i < measurements.__len__():
        tmp = measurements[i: i + 100000]
        Measurements.objects.bulk_create(tmp)
        i += 100000
    # return data_par_obj_all


def find_parameter(obj, data_par_obj_all):
    for data in data_par_obj_all:
        if data.par_rollennr != obj['parameter']['par_rollennr']:
            continue
        if data.par_definition_id != obj['parameter']['par_definition_id']:
            continue
        if data.par_name != obj['parameter']['par_name']:
            continue
        return data

Answer 1

您需要使用比列表更好的数据结构，以防止重复。

from itertools import zip_longest

def handle_parameters_upload(request, file):
    wb = openpyxl.load_workbook(file, read_only=True)
    first_sheet = wb.get_sheet_names()[0]
    ws = wb.get_sheet_by_name(first_sheet)

    recipe, _ = Recipe.objects.get_or_create(par_recipe=ws["B2"].value)

    line, _ = Line.objects.get_or_create(par_machine=ws["C2"].value)

    order, _ = Order.objects.get_or_create(par_fa=ws["D2"].value)

    # Clear existing data
    Measurements.objects.all().delete()
    Parameter.objects.all().delete()

    parameter_data = set()
    duplicate_measurement_parameter = {}
    measurement_data = []

    for row in ws.iter_rows(row_offset=1):
        parameter = (row[5].value, row[6].value, row[7].value,)

        if parameter in parameter_data:
            duplicate_measurement_parameter[row[8].value] = {
               'par_rollennr': parameter[0],
               'par_definition_id': parameter[1],
               'par_name': parameter[2],
            }
        # Add the parameter
        parameter_data.add(parameter)
        # par_rollennr, par_definition_id, par_name, measurement: par_value   
        measurement_data.append(row[8].value)

    # Bulk create data
    parameters = Parameter.objects.bulk_create([
        Parameter(
            id_parameter=index, 
            par_rollennr=p_data[0], 
            par_definition_id=p_data[1],
            par_name=p_data[2],
       ) for index, p_data in enumerate(parameter_data) 
    ])
    Measurements.objects.bulk_create([
        Measurements(
            line=line,
            order=order,
            recipe=recipe,
            par_value=m_data,
            parameter=parameter or Parameter.objects.filter(**duplicate_measurement_parameter.get(m_data, {})).first(),
        ) 
        for parameter, m_data in zip_longest(parameters, measurement_data)
    ])
    return True

如何避免数据库中记录的重复？

1 个答案: