ElasticSearch how it works

I want to understand how ElasticSearch works. I know it uses an inverted index structure, but reading from some articles there are tons of different ways to build such a structure.

My questions are as follows:

1. What information is saved when the data is loaded by ElasticSearch? (word frequency, position, ...)
2. There is a file on the path I can check to see the saved structure of the map? (something in the data folder maybe?)
3. Is there any difference with the use of numerical fields? (From the official documentation it seems to use a different technique, i.e. BKD trees)

Anyone knows a detailed book that can help me understand how the engine works? I mean, what happens when i load a json file and what happens when I run a query.

The best books of Elasticsearch are:

• Gheorghe R., Hinman M.L., Russo R. Elasticsearch in Action. Manning, 2016.
• Gormley G., Tong Z. Elasticsearch: The Definitive Guide. O’Reilly, 2015.
• Konda M. Elasticsearch in Action, 2nd ed. Manning, 2023.
• Paro A. Elasticsearch 8.x Cookbook, 5th ed. Packt, 2022

See almost full list of books about Elasticsearch and why these 4 books are best