Table A
Id T ci

1 I am XXXX
2 I'm XXX
3 who are you XXXXX
4 who are you XXXX
5 I'm XXX
I am in XXXX


Above table data structure is
Article 10 million

Evaluate a method or an SQL can directly deleted directly T this field the same data only keep a become

1 I am XXXX
3 who are you XXXXX
5 I am XXXX


Are there any simple way the forehead efficiency does not need to pursue, and server hanging inside, anyway, and don't crash

CodePudding user response:

This is the last post told you direct reason of es in

Morning, since the contact with these things in the evening you will have to es, on the contrary play Nlp yourself from scratch, we do not recommend, Nlp is, after all, so-called "pearl in the crown to artificial intelligence", can oneself study, we do not recommend project to do, the pearl is not who want to pick can pick

Let's look at es how

https://blog.csdn.net/truenaruto/article/details/81120196

Es get things on the first, and then pick a pearl, let alone es is not technology, let alone get es, NLP after waste, even when you are going to NLP, es are infrastructure waste drops,

CodePudding user response:

No build tutorial yao didn't contact with this see if baidu need to configure the JAVA environment

CodePudding user response:

refer to the second floor ali hat response:

not build tutorial that no contact with this see if baidu need to configure the JAVA environment

Development is really good, only docker installation docker and then install the somebody else has been built in the docker into a mirror, we only use the mirror to develop, as for the deployment, operations of other people dedicated lane,

So I posted before said such a thing is best not to let the programmers involved, should let the product manager and technical manager intervention,
How the product positioning, how to design the target node, how to set up a milestone for products to do
What kind of, how to do, how to build, technical manager,

Programmers put products and technical manager, live ops are dry, hard also thankless, after all, to the back of his head, not in the position, consideration, can't plug in

CodePudding user response:

The docker, the installation es
https://www.cnblogs.com/powerbear/p/11298135.html

CodePudding user response:

The DELETE FROM the text WHERE id NOT IN
(SELECT dt. Mins_id FROM
(SELECT MIN (id) AS mins_id FROM text GROUP BY text)
Dt);


This is very good, looking for a few days ago, requirements and you are same

CodePudding user response:

If you don't es, the feeling is too complex,
So, I can only suggest that you, the background to run a job,
Then he sorting data for you, every day in a new table,

CodePudding user response:

According to the above job,
On the principle of actually es are also help you to do this,
Nothing more than what you put in another table, in the es,
But by the lexical analysis of the es, help you use the inverted index method to store data,

CodePudding user response:

In addition, I don't recommend deleting data directly,
You most as a data filter, but there is no special situation, it is best not to delete the original data,

CodePudding user response:

Don't consider efficiency, with a trigger is quite convenient