Python 比较列表中的两个二维列表，并打印不同的行。但是没有一列

以下是：

gnucashumsaetze = [
 ['2020-11-27', 'Essen', '4.53'],
 ['2020-11-27', 'Essen', '10.67'],
 ['2020-11-30', 'Essen', '4.80'],
 ['2020-11-30', 'Lebensmittel', '2.78'],
 ['2020-11-30', 'Essen', '2.31'],
 ['2020-11-30', 'Kosmetik', '5.58'],
 ['2020-12-01', 'Essen', '11.23'],
]

onlineumsaetze = [
['2020-11-27', 'EDEKA ERNST HAUPTBAHNH  / MUENCHEN', '4.53']
['2020-11-27', 'Netto Marken-Discount  / Ingolstadt', '10.67']
['2020-11-30', 'MUELLER GMBH & CO.KG  / NUERNBERG', '4.80']
['2020-11-30', 'Netto Marken-Discount  / Frankfurt', '2.31']
['2020-11-30', 'Rossmann 2380  / Ingolstadt', '5.58']
['2020-11-30', 'ALIEXPRESS.COM  / Luxembourg', '22.46']
['2020-12-01', 'EDEKA BRAUN  / INGOLSTADT', '11.23']
['2020-12-02', 'EDEKA ERNST HAUPTBAHNH  / MUENCHEN', '7.03']
]

我想比较两个2d列表并输出不同的结果。但不应比较第二列（第[1]行）。像这样：

['2020-11-30', 'ALIEXPRESS.COM  / Luxembourg', '22.46']
['2020-12-01', 'EDEKA BRAUN  / INGOLSTADT', '11.23']
['2020-12-02', 'EDEKA ERNST HAUPTBAHNH  / MUENCHEN', '7.03']

我已经尝试过的是这个；不幸的是，这是一场灾难：

fehlende_rows = (set((row[0] for row in onlineumsaetze),(row[2] for row in onlineumsaetze)) - set((row[0] for row in gnucashumsaetze),(row[2] for row in gnucashumsaetze)))
print(fehlende_rows)

---

我发现先写出完整的循环，然后尽可能地把它浓缩成一个列表理解，这真的很有帮助

最好的方法可能是迭代

gnucashumsaetze

并创建一个字符串->集合字典，其中日期作为键，数字作为集合的元素

gnucashumsaetze_dict = {}
for g in gnucashumsaetze:
    date, val = g[0], g[2]
    # Maybe you want to do val = float(g[2]) instead?
    if date not in gnucashumsaetze_dict:
        gnucashumsaetze_dict[date] = set()
    gnucashumsaetze_dict[date].add(val)

gnucashumsaetze_dict

现在是：

{'2020-11-27': {'10.67', '4.53'},
 '2020-11-30': {'2.31', '2.78', '4.80', '5.58'},
 '2020-12-01': {'11.23'}}

[['2020-11-30', 'ALIEXPRESS.COM  / Luxembourg', '22.46'],
 ['2020-12-02', 'EDEKA ERNST HAUPTBAHNH  / MUENCHEN', '7.03']]

然后，迭代

onlineumsaetze

中的每一行，并仅在满足所需条件时将其附加到新列表中

new_onlineumsaetze = []
for o in onlineumsaetze:
    date, val = o[0], o[2]
    # if date is not in gnucashumsaetze_dict, return empty set
    vals = gnucashumsaetze_dict.get(date, set()) 
    if val not in vals:
        new_onlineumsaetze.append(o)

new_onlineumsaetze

现在是：

{'2020-11-27': {'10.67', '4.53'},
 '2020-11-30': {'2.31', '2.78', '4.80', '5.58'},
 '2020-12-01': {'11.23'}}

[['2020-11-30', 'ALIEXPRESS.COM  / Luxembourg', '22.46'],
 ['2020-12-02', 'EDEKA ERNST HAUPTBAHNH  / MUENCHEN', '7.03']]

跳过

['2020-12-01'，'EDEKA BRAUN/INGOLSTADT'，'11.23']

行，因为

gnucashumsaetze

有一个

['2020-12-01'，'Essen'，'11.23']

既然您已经将其作为常规for循环编写，那么就更容易将其浓缩为一个列表

new_onlineumsaetze = [o for o in onlineumsaetze if o[2] not in gnucashumsaetze_dict.get(o[0], set())]

---

为了解决这个问题，我将使用列表理解

首先仅使用column0和column2创建两个集

gnucashumsaetze_set = set([(row[0], row[2]) for row in gnucashumsaetze])
onlineumsaetze_set = set([(row[0], row[2]) for row in onlineumsaetze])

然后我们得到这两个集合的差

diff_ = onlineumsaetze_set.difference(gnucashumsaetze_set)

对于最终结果，我们在onlineumsaetze中查找与我们获得的数据在column0和column2中匹配的行

res = [row for row in onlineumsaetze if (row[0], row[2]) in diff_]

print(res)

结果

[['2020-11-30', 'ALIEXPRESS.COM  / Luxembourg', '22.46'], ['2020-12-02', 'EDEKA ERNST HAUPTBAHNH  / MUENCHEN', '7.03']]

---

是否允许您更改数据的结构？比如把这些列表转换成dicts？你说的“输出不同的”是什么意思？您的意思是，仅当

gnucashumsaetze

第三列中不包含具有相同日期和数值的行时，您才想从

onlineumsaetze

输出一行吗？@Diptangsu Goswami，很遗憾，不是。此数据是从HTML网页复制的。@Pranav Hosangadi，是的，确切地说，在第一列和第三列。