dashjay commented on PR #3269:
URL: https://github.com/apache/kvrocks/pull/3269#issuecomment-3712889309
> @dashjay @PragmaTwice @git-hulk Good point on the marker.关于标记物这一点说得很好。 I
agree that using value prefixes (even multiple bytes) is not a collision free
solution because the hash field value is opaque binary
data.我同意使用值前缀（即使是多个字节）并不是一个无冲突的解决方案，因为哈希字段值是不透明的二进制数据。 Based on your experience
what would you suggest?根据您的经验，您会提出什么建议？ Could we do a major release where we
introduce a new format for hash fields that has a metadata object encoded
within the hash field value?我们能否发布一个重大版本，引入一种新的哈希字段格式，将元数据对象编码到哈希字段值中？
@ltagliamonte-dd I had a good dream lol.
I used to use encoded key/value to implement overlay key-value DB like
distribute MVCC key value store, we encoded the timestamp from TSO into the
user key for MVCC, we use scan for user get lol, sound very expensive, but it
works, so we can store the value in key `{user_key}:{TTL}`，if there is no TTL
keep it maxint64 can make it higher in lexicographical order. If we did like
this every get will be scan, need to create new iter.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
