MySQL INSERT ...使用django 1.4进行DUPLICATE KEY UPDATE批量插入

mmo*_*udd 3 python mysql django

我有问题搞清楚MySQL INSERT ...用django 1.4进行DUPLICATE KEY UPDATE.

我尝试插入记录的表有一个2列(复合)唯一键.我收到的记录来自第三方来源,除了制作唯一键集的字段外,值会随时间变化.我一次收到1~5k的记录,需要

目前我正在使用Model.objects.bulk_create进行批量插入,性能非常惊人,因为无论记录集有多大,它通常会发出一个查询.但是,由于我的记录可能会在第三方结束时随时间发生变化,因此我需要在记录集上执行MySQL INSERT ... ON DUPLICATE KEY UPDATE查询.

我打算编写原始SQL语句并使用类似的方式执行:

sql = "MySQL INSERT ... ON DUPLICATE KEY UPDATE"

raw_insert(sql)

def raw_insert(sql):
    from django.db import connection, transaction
    cursor = connection.cursor()

    # Data modifying operation - commit required
    cursor.execute(sql)
    transaction.commit_unless_managed()

    return 1
Run Code Online (Sandbox Code Playgroud)

想知道我的问题是否有更好的解决方案.另外,我如何清理原始插入的字段值?

mmo*_*udd 9

所以我创建了一个自定义管理器 这是经理:

class BulkInsertManager(models.Manager):
    def _bulk_insert_or_update(self, create_fields, update_fields, values):

        from django.db import connection, transaction
        cursor = connection.cursor()

        db_table = self.model._meta.db_table

        values_sql = []
        values_data =[]

        for value_lists in values:
            values_sql.append( "(%s)" % (','.join([ "%s" for i in range(len(value_lists))]),) )
            values_data.extend(value_lists)

        base_sql = "INSERT INTO %s (%s) VALUES " % (db_table, ",".join(create_fields))

        on_duplicates = []

        for field in update_fields:
            on_duplicates.append(field + "=VALUES(" + field +")")

        sql = "%s %s ON DUPLICATE KEY UPDATE %s" % (base_sql, ", ".join(values_sql), ",".join(on_duplicates))

        cursor.executemany(sql, [values_data])
        transaction.commit_unless_managed()
Run Code Online (Sandbox Code Playgroud)

和样本模型:

class User_Friend(models.Model):
    objects = BulkInsertManager() # assign a custom manager to handle bulk insert

    id = models.CharField(max_length=255)
    user = models.ForeignKey(User, null=False, blank=False)
    first_name = models.CharField(max_length=30)
    last_name = models.CharField(max_length=30)
    city = models.CharField(max_length=50, null=True, blank=True)
    province = models.CharField(max_length=50, null=True, blank=True)
    country =  models.CharField(max_length=30, null=True, blank=True)
Run Code Online (Sandbox Code Playgroud)

示例实现:

def save_user_friends(user, friends):
    user_friends = []
    for friend in friends:

        create_fields = ['id', 'user_id', 'first_name', 'last_name', 'city', 'province', 'country']
        update_fields = ['first_name', 'last_name', 'city', 'province', 'country']

        user_friends.append(
            [
                str(user.id), 
                str(friend['id']),
                friend['first_name'],
                friend['last_name'],
                friend['city'],
                friend['province'],
                friend['country'],
            ]
        )

    User_Friend.objects._bulk_insert_or_update(create_fields, update_fields, user_friends)
Run Code Online (Sandbox Code Playgroud)

这是要点.