简体繁体 English

Elasticsearch：是新类型还是新索引？

[英]Elasticsearch: new type or new index?

原文 2018-02-10 12:18:46 0 1 elasticsearch/ kibana

I am doing a test, gathering tweets from the Twitter API, and index some of the information in Elasticsearch. 我正在做一个测试，从Twitter API收集推文，并为Elasticsearch中的某些信息建立索引。

As we all know, Elasticsearch doesn't handle well array of objects . 众所周知，Elasticsearch不能很好地处理对象数组 。 And I need those for some analysis. 我需要进行一些分析。 Examples of arrays of objects from a tweet could be: 来自推文的对象数组的示例可能是：

List of urls contained in the text, with information about domain, or other things 文本中包含的网址列表，以及有关域或其他内容的信息
List of mentions contained in the text, with information about the user (like join date, user_id) 文本中包含的提及列表，以及有关用户的信息（例如加入日期，user_id）
Tokens and semantyc analysis from the text. 文本中的标记和语义分析。

Kibana doesn't handle well, as well, array of objects and it makes a nightmare to visualize these objects. Kibana也无法很好地处理对象数组，并且将这些对象可视化是一场噩梦。 So I was thinking to flat those objects in a new index or a new type of the same index, putting in some information abbout the origin tweet . 所以我想平这些对象在一个新的索引或新类型相同指数，投入abbout起源的一些信息tweet 。

What is better? 什么是更好的？ a new index? 新索引？ a new type? 一种新类型？ and why? 为什么呢？ how can I make a visualization on Kibana that takes in consideratio the id from another type/index? 如何在Kibana上进行可视化处理，考虑其他类型/索引的ID？ which is better? 哪个更好？

1 个解决方案

Mapping types are deprecated in elasticsearch 6.x and will be removed in 7.x. 映射类型在elasticsearch 6.x中已弃用，并将在7.x中删除。 So I guess a new index is the way to go. 因此，我想一个新的索引就是要走的路。