类似： https://leetcode.com/problems/lru-cache/

窃以为，考虑这个问题的满足(satisfaction)问题版本：判断是否存在任意一个行的组合，满足所有列的约束条件。可以很容易的 reduce 成 SAT 问题，所以可以肯定是一个 NP 问题。如此一来，基本只能靠一些 heuristics 来剪枝搜索空间了。

也说不上是错误，更多是自然语言的局限性。如果从逻辑学来看，and 满足交换律，A and B = B and A ；从计算角度看，A and B 是先判断 A 在判断 B 。如果不比较几个版本，你也看不出作者到底想说什么语义。

所以我说这书语序有问题，ppt 就是你说的语序。

如果 Tk 还没有写完，Ti 的 validation 无法判断“Tk 所写的数据项集与 Ti 所读数据项集不相交”。

第 2 点其实就是说两个事务如果运行时间有重合的时候怎么办。
你看看第七版原版说得是不是合理。ppt 第 50 页： https://www.db-book.com/db7/slides-dir/PDF-dir/ch18.pdf
（你会发现这书说得颠三倒四，ppt 和第 6 版说得好像是同一件事情，但是语序换了一下可以让读者理解成完全不同的事情。）

最常用的就是η-reduction，λ x. f x = f

select * from world where continent = "Asia" order by population desc limit 3 offset 2;

C 里存 A, B 的 reference

>>> s = 'babble'
>>> re.sub('(?<!^)' + s[0], '*', s)
'ba**le'

估计 lz 没有发现 linked list 和 tree 的联系。

@aijam 把这一行加入.bashrc。忘 alias 了，就输入 alias 查询下。

alias alias='alias | less'

In Assembly or C, type determines the interpretation of a sequence of memory and the operations that can be performed on it. For example, if you interpret 4 bytes of data as int, you can perform arithmetic operations with it; but if you interpret the same 4 bytes as char[4], you will be able to iterate over it with some indices.
Such idea can be found in many other languages. Another example is when you do "type casting", you are basically indicating the compiler to treat the same sequence of bytes with a different interpretation, such that a different set of operations can be applied.
So, how can you create your own type, i.e. define your own interpretation and the set of operations on a piece of data? That's where classes come into the picture. In general, you can consider a class as a recipe of a user-defined type, in contrast to the built-in types which are already predefined for you.
Surely, types and classes are much more profound concepts, especially when putting in the context of different languages. If you want to dive deeper on this topic, I think you should learn a little more about types and "type class" in Haskell, and maybe try to find the connection between "type class" in Haskell and "abstract class"/"interface" in the OO world. Also, compare them with "struct" in C/C++ and see how those languages provide their own mechanisms for abstraction.

觉得不值就卖给别人呗

@wutiantong 其实我也这么觉得，只是看看 greedy 分组啥效果 :D

@aijam 只能蒙着眼瞎写猜测 greedy 了。。。