NamedTuple

資料分析或機器學習領域，用好 NamedTuples 會寫出可讀性更強、更易於維護的程式碼。

做特徵工程的時候，如果把特徵扔到一個list當中，以便日後取用，但是取用的時候，難免出現整數索引，程式碼可讀性差，所以使用NamedTuple，避免出現整數索引，希望是直接按屬性索引：

from collections import namedtuple

#建立一個帶有 14 個屬性，名字為 Person 的 NamedTuple 例項 Person
Person = namedtuple('Person',['id','age','height','name','address
 
','province','city','town','country','birth_address','father_name', 'monther_name','telephone','emergency_telephone'])
# 呼叫例項 Person，建立一個 id=10086 的 Person 物件
a = ['']*11
Person(10086,19,'xiaoming',*a)

output:
Person(id=10086, age=19, height='xiaoming', name='', address='', province='', city='', town='', country=''
 
, birth_address='', father_name='', monther_name='', telephone='', emergency_telephone='')

假設有個任務，再有老資料的情況下，有了一份新資料，現在要比較，哪些人的居住地址（對應欄位 address）、聯絡電話（對應欄位 telephone）、出生地資訊（對應欄位 birth address）發生了變化，統計出這些人。

使用NamedTuple方法：

def update_persons_info(old_data,new_data):
    changed_list = []
    for line in
 
 new_data:
        new_props = line.split() 
        new_person = Person(new_props) # new_props 與 Person 引數卡對好
        for old in old_data: 
            old_props =  old.split() 
            old_person = Person(old_props)
            if old_person.id != new_person.id: 
                changed_list.append(old_person.id)
            elif old_person.address != new_person.address:
                changed_list.append(old_person.address)
            elif old_person.birth_address != new_person.birth_address: 
                changed_list.append(old_person.birth_address)
    return changed_list

但是在帶來這樣的遍歷同時，也帶來一個問題，NamedTuple 建立後，它的屬性取值不允許被修改，也就是屬性只能是可讀的，就看怎麼用了。

from collections import namedtuple

#建立一個帶有 14 個屬性，名字為 Person 的 NamedTuple 例項 Person
Person = namedtuple('Person',['id','age','height','name','address','province','city','town','country','birth_address','father_name', 'monther_name','telephone','emergency_telephone'])
# 呼叫例項 Person，建立一個 id=10086 的 Person 物件
a = ['']*11
xiaoming = Person(10086,19,'xiaoming',*a)
print(type(xiaoming))
xiaoming.age = 20

output:
<class '__main__.Person'>
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-5-dabf7d8e12b1> in <module>()
      7 xiaoming = Person(10086,19,'xiaoming',*a)
      8 print(type(xiaoming))
----> 9 xiaoming.age = 20

AttributeError: can't set attribute

Counter

主要用於統計中的計數，使用 Counter，期待能寫出更加簡化的程式碼

from collections import Counter

# 統計出現次數
freq = [3, 8, 3, 10, 3, 3, 1, 3, 7, 6, 1, 2, 7, 0, 7, 9, 1, 5, 1, 0]
Counter(freq).most_common()

output:
[(3, 5),
 (1, 4),
 (7, 3),
 (0, 2),
 (8, 1),
 (10, 1),
 (6, 1),
 (2, 1),
 (9, 1),
 (5, 1)]

並且，是按照頻數由高到低排序的，牛的

使用 Counter 能快速統計，一句話中單詞出現次數，一個單詞中字元出現次數。如下所示：

text = """
def update_persons_info(old_data,new_data):
    changed_list = []
    for line in new_data:
        new_props = line.split() 
        new_person = Person(new_props) # new_props 與 Person 引數卡對好
        for old in old_data: 
            old_props =  old.split() 
            old_person = Person(old_props)
            if old_person.id != new_person.id: 
                changed_list.append(old_person.id)
            elif old_person.address != new_person.address:
                changed_list.append(old_person.address)
            elif old_person.birth_address != new_person.birth_address: 
                changed_list.append(old_person.birth_address)
    return changed_list"""
Counter(text).most_common()

output:
[(3, 5),
 (1, 4),
 (7, 3),
 (0, 2),
 (8, 1),
 (10, 1),
 (6, 1),
 (2, 1),
 (9, 1),
 (5, 1)]

text = """
def update_persons_info(old_data,new_data):
    changed_list = []
    for line in new_data:
        new_props = line.split() 
        new_person = Person(new_props) # new_props 與 Person 引數卡對好
        for old in old_data: 
            old_props =  old.split() 
            old_person = Person(old_props)
            if old_person.id != new_person.id: 
                changed_list.append(old_person.id)
            elif old_person.address != new_person.address:
                changed_list.append(old_person.address)
            elif old_person.birth_address != new_person.birth_address: 
                changed_list.append(old_person.birth_address)
    return changed_list"""
Counter(text).most_common()
text = """
def update_persons_info(old_data,new_data):
    changed_list = []
    for line in new_data:
        new_props = line.split() 
        new_person = Person(new_props) # new_props 與 Person 引數卡對好
        for old in old_data: 
            old_props =  old.split() 
            old_person = Person(old_props)
            if old_person.id != new_person.id: 
                changed_list.append(old_person.id)
            elif old_person.address != new_person.address:
                changed_list.append(old_person.address)
            elif old_person.birth_address != new_person.birth_address: 
                changed_list.append(old_person.birth_address)
    return changed_list"""
Counter(text).most_common()
[(' ', 182),
 ('e', 45),
 ('d', 42),
 ('s', 40),
 ('n', 38),
 ('o', 36),
 ('r', 33),
 ('p', 31),
 ('_', 30),
 ('l', 24),
 ('a', 23),
 ('i', 21),
 ('t', 16),
 ('\n', 15),
 ('.', 14),
 ('w', 9),
 ('(', 8),
 (')', 8),
 ('h', 8),
 ('=', 8),
 ('f', 7),
 (':', 6),
 ('c', 5),
 ('g', 5),
 ('P', 3),
 ('!', 3),
 ('b', 3),
 ('u', 2),
 (',', 1),
 ('[', 1),
 (']', 1),
 ('#', 1),
 ('與', 1),
 ('參', 1),
 ('數', 1),
 ('卡', 1),
 ('對', 1),
 ('好', 1)]

DefaultDict

DefaultDict 能自動建立一個被初始化的字典，也就是每個鍵都已經被訪問過一次。

如何建立預設初始化某型別的字典值

from collections import defaultdict

# 建立一個字典值型別為 int 的預設字典：
dict_1 = defaultdict(int)
# 建立一個字典值型別為 list 的預設字典：
dict_2 = defaultdict(list)
dict_2

output:
defaultdict(list, {})

s = 'from collections import defaultdict'
for index,i in enumerate(s):
    dict_2[i].append(index)
print(dict_2)

output:
defaultdict(<class 'list'>, {'f': [0, 26], 'r': [1, 21], 'o': [2, 6, 13, 20], 'm': [3, 18], ' ': [4, 16, 23], 'c': [5, 10, 33], 'l': [7, 8, 29], 'e': [9, 25], 't': [11, 22, 30, 34], 'i': [12, 17, 32], 'n': [14], 's': [15], 'p': [19], 'd': [24, 31], 'a': [27], 'u': [28]})