Python for 循环在迭代 CSV 行时不合理地中途停止

Question

我正在处理一个CSV文件来分析讲座反馈数据，格式如下

"5631","18650","10",,,"2015-09-18 09:35:11"
"18650","null","10",,,"2015-09-18 09:37:12"
"18650","5631","10",,,"2015-09-18 09:37:19"
"58649","null","6",,,"2015-09-18 09:38:13"
"45379","31541","10","its friday","nothing yet keep it up","2015-09-18 09:39:46"

我正在尝试删除不良数据。只有具有“id1”、“id2”和另一个对应的“id2”、“id1”的数据条目才被视为有效。

我正在使用嵌套循环来尝试为每行查找匹配的条目。然而，外循环似乎无缘无故地中途停止了。这是我的代码

class Filter:
    file1 = open('EncodedPeerInteractions.FA2015.csv')
    peerinter = csv.reader(file1,delimiter=',') 
    def __init__(self):
        super()

    def filter(self):
        file2 = open('FilteredInteractions.csv','a')
        for row in self.peerinter:
            print(row)
            if row[0] == 'null' or row[1] == 'null':
                continue
            id1 = int(row[0])
            id2 = int(row[1])
            for test in self.peerinter:
                if test[0] == 'null' or test[1] == 'null':
                    continue
                if int(test[0]) == id2 and int(test[1]) == id1:
                    file2.write("\n")
                    file2.write(str(row))
                    break
        file2.close()

我尝试使用 pdb 单步执行代码，前几个循环一切正常，然后突然跳转到 file2.close() 并返回。该程序确实打印出了一些有效的条目，但这还不够。

我测试了 csv 文件并将其正确加载到内存中，其中包含超过 18000 个条目。我使用 print 进行了测试，但它给出了相同的结果，因此附加文件没有任何问题。

编辑

现在我明白问题所在了。如this question说，当有匹配时我会中断，但当没有匹配时，内部循环将消耗所有文件而不重置它。当它返回到外循环时，它就结束了。我应该将其放入列表或让它重置。

Answer 1

你让这种方式变得更加复杂。

给定:

$ cat /tmp/so.csv
"5631","18650","10",,,"2015-09-18 09:35:11"
"18650","null","10",,,"2015-09-18 09:37:12"
"18650","5631","10",,,"2015-09-18 09:37:19"
"58649","null","6",,,"2015-09-18 09:38:13"
"45379","31541","10","its friday","nothing yet keep it up","2015-09-18 09:39:46"

您可以使用 csv 和过滤器来获取您想要的内容:

>>> with open('/tmp/so.csv') as f:
...    list(filter(lambda row: 'null' not in row[0:2], csv.reader(f)))
... 
[['5631', '18650', '10', '', '', '2015-09-18 09:35:11'], 
 ['18650', '5631', '10', '', '', '2015-09-18 09:37:19'], 
 ['45379', '31541', '10', 'its friday', 'nothing yet keep it up', '2015-09-18 09:39:46']]