使用executemany将CSV数据插入Django模型

Question

我有一个包含大约90万行和13列的csv文件，前28445行一切正常，但之后就出现了错误。

编程错误

异常信息：格式字符串的参数不够

我尝试通过打印行来检查列是否有问题，但看起来没有什么异常。

['印度', '5', '第一次当妈妈', '帕特纳', 'A2', '品牌驱动因素', '', '', '很难找到', '', '', '1', '0到12个月']

def upload(request):
    if request.method == 'POST':
        cursor = connection.cursor()
        query = ''' INSERT INTO johnson_jnjusage (country,no_of_people_house,nursing_cnt,city,sec,bucket,category1,category2, final_category, responders, usageFrequency, base, child_age_group) 
                    VALUES(%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s) '''
        x=[]
        reader = csv.reader(request.FILES['csvData'],delimiter=',')
        i = 0
        for row in reader:
            x.append(tuple(row))
            if i>=5000:
                cursor.executemany(query,tuple(x))
                transaction.commit()
                x=[]
                i=0

            i = i+1

        cursor.close()
        return HttpResponse( docfile.name + "'s data inserted into database successfully")

如果你们能帮我解决这个问题，我将非常感激。

Answer 1

我明白了。我只是检查了一下那一行的大小，然后对那一行进行了处理。同时，我发现有一些字符，比如“\x00”，所以我用正则表达式把它们去掉了。

def upload(request):
    start_time = time.time()
    print start_time
    if request.method == 'POST':
        cursor = connection.cursor()    
        x=[]
        docfile = request.FILES['csvData']

        reader = csv.reader(request.FILES['csvData'],delimiter=',')

        to_create = []
        for i, row in enumerate(reader):
            if len(row) != 13:
                reader.next()
                continue

            j = JnJUsage()
            j.country =row[0]
            j.no_of_people_house = row[1]
            j.nursing_cnt = row[2]
            j.city = row[3]
            j.sec = row[4]
            j.bucket = row[5]
            j.category1 = row[6]
            j.category2 = row[7]
            j.final_category = re.sub(r'[\x00-\x08\x0b\x0c\x0e-\x1f\x7f-\xff]', '', row[8])
            j.responders = row[9]
            j.usageFrequency = row[10]
            j.base = row[11]
            j.child_age_group = row[12]

            to_create.append(j)

            # If 900k is too much then you could consider something like this
            if i % 10000 == 0:
                JnJUsage.objects.bulk_create(to_create)
                to_create = []
        JnJUsage.objects.bulk_create(to_create)

        return HttpResponse( docfile.name + "'s data inserted into database successfully")

感谢Dmitry Mikhaylov和Nathaniel的帮助。

Answer 2

当然，你可以使用批量创建（bulk_create）。它的写法大概是这样的：

def upload(request):
    if request.method == 'POST':
        cursor = connection.cursor()
        x=[]
        reader = csv.reader(request.FILES['csvData'],delimiter=',')
        i = 0
        for row in reader:
            obj = new MyObject()
            obj.country = row[0]
            obj.city = row[3]
            ...
            x.append(obj)
            if i>=5000:
                MyObject.objects.bulk_create(x)
                x=[]
                i=0

            i = i+1
        return HttpResponse( docfile.name + "'s data inserted into database successfully")

你可以在文档中找到更多关于bulk_create的信息。

Answer 3

你可以这样做。

#models.py
class JNJUsage(models.Model):
    ...

# views.py (where ever def upload is)
to_create = []

for i, row in enumerate(reader):
    j = JNJUsage()
    j.country = row['country']
    j.no_of_people_house = row['no_of_people_house']
    j.nursing_cnt = row['nursing_cnt']
    j.city = row['city']
    j.sec = row['sec']
    j.bucket = row['bucket']
    j.category1 = row['category1']
    j.category2 = row['category2']
    j.final_category = row['final_category']
    j.responders = row['responders']
    j.usageFrequency = row['usageFrequency']
    j.base = row['base']
    j.child_age_group = row['child_age_group']

    to_create.append(j)

    # If 900k is too much then you could consider something like this
    if i % 10000 == 0:
        JNJUsage.objects.bulk_create(to_create)
        to_create = []

# Clean up the rest
JNJUsage.objects.bulk_create(to_create)