不想用mysql的自增id(造資料的時候會很麻煩),而且tiDB也不支援.

而且主鍵希望和資源的業務屬性是可對應的.

比如裝置就是域+ip,這樣如果刪除了再新增,主鍵不會改變.

直接用uuid3,32位元組的字串有點太長,16byte的bytes操作又麻煩, 128bit的int又超過了mysql uint64的限制(拆2位元組太麻煩)

搜尋了一下uuid的壓縮方法,發現base64編碼還是比較靠譜的.

使用base64編碼可以壓縮到22byte, 不過python的base64函式是需要補齊4位的.

檢視文件,還有b85編碼,可以進一步壓到20byte,而且b85encode/b85decode是預設不填充的.

ok,就它了

import uuid
import hashlib

name = 'test_nameqqqwwe1234567890'

#namespace = uuid.NAMESPACE_URL
namespace = uuid.NAMESPACE_DNS

print(uuid.uuid1())
print(uuid.uuid3(namespace,name))
print(uuid.uuid3(namespace,name).int)
print(uuid.uuid3(namespace,name).bytes)
print(uuid.uuid3(namespace,name).hex)
print(uuid.uuid4())
print(uuid.uuid5(namespace,name))
print(hashlib.md5(name.encode(encoding='UTF-8')).hexdigest())

#測試將字串轉為uuid物件
print(uuid.UUID('7e9fc1763b19394a953083391161f8e9'))

#測試用base64來壓縮uuid
import base64

print('test str base64')
def safe_base64_encode(s):
    b = base64.b64encode(s.encode('utf-8'))#因為python3.x中字元都為unicode編碼，而b64encode函式的引數為byte型別，所以必須先轉碼    
    #b= base64.b64encode(s)
    bstr_tmp = str(b,'utf-8') #把byte型別的資料轉換為utf-8的資料    
    print(bstr_tmp)
    b_str= bstr_tmp.strip(r'=+') #用正則把 = 去掉    
    return b_str

def safe_base64_encodebyte(bs):  #b64encode只能接收byte
    #b = base64.b64encode(s.encode('utf-8'))#因為python3.x中字元都為unicode編碼，而b64encode函式的引數為byte型別，所以必須先轉碼    
    b= base64.b64encode(bs)
    bstr_tmp = str(b,'utf-8') #把byte型別的資料轉換為utf-8的資料    
    print(bstr_tmp)
    b_str= bstr_tmp.strip(r'=+') #用正則把 = 去掉    
    return b_str
    
def safe_base64_decode(s):   #b64decode可以接收字串或者byte
    import math
    padlen=math.ceil(len(s)/4)*4
    print(padlen)
    pstr=ten64.ljust(padlen,'=')
    return pstr

s = "binarybstr\x00string"
safe_b = safe_base64_encode(s)
print (safe_b)


print("test base64")
en64=base64.b64encode(uuid.uuid3(namespace,name).bytes)  #b64encode只能接收byte
#en64=base64.b64encode(uuid.uuid3(namespace,name).hex.encode('utf-8'))  #b64encode只能接收byte. 這樣更長了
print(en64)
#trime尾部的=
#ten64=str(en64,'utf-8').strip(r'=+')
ten64=safe_base64_encodebyte(uuid.uuid3(namespace,name).bytes)
print(ten64)
#恢復時補全尾部的=
ren64=safe_base64_decode(ten64)
#ren64=ten64.ljust(24,'=')      #uuid16byte *4/3=22byte,補全為4的倍數24
d64=base64.b64decode(ren64)    #b64decode可以接收字串或者byte
print(d64)


print('test base85')
en85=base64.b85encode(uuid.uuid3(namespace,name).bytes)    #20位，而且不用考慮==
print(en85)
de85=base64.b85decode(en85)
print(de85)
 

test base85
b'FcW&B=DR*3jUe2byo||5'
b'0\x13z\xa2\xe6\xbb>"\x8d \xdc\x9a\xbc\x8c\xc9G'

不過, b85不是urlsafe的

如果需要在瀏覽器裡面直接用,還是用b64更保險