Mentioned earlier: Insert ... select ... where not exist
In fact, it also seems to be implemented: when inserting data, do not insert if there are duplicates.
See also: http://snowolf.iteye.com/blog/1568926
and insert IGNORE into almost, but the former can add their own conditions, the latter only according to the primary key to determine whether to repeat.
MySQL Some development specifications:
Http://www.2cto.com/database/201403/285170.html
1. The true meaning of MYSQL sql_no_cache Http://www.dewen.org/q/5149/Mysql
The result is not cached, but the query is still cached.
If you want to re-test, please execute "FLUSH query cache" before querying, and clear the query cache
2.INSERT ... On DUPLICATE KEY UPDATE
It is possible to implement the original key record if there is an update, not present the inserted.
(REPLACE into will delete the original duplicate record, then insert)