elastic / elasticsearch

Free and Open Source, Distributed, RESTful Search Engine
https://www.elastic.co/products/elasticsearch
Other
69.67k stars 24.66k forks source link

Path hierarchy tokenizer doc could be a bit more helpful #17138

Closed kix closed 5 years ago

kix commented 8 years ago

Elasticsearch version: current

JVM version: irrelevant

OS version: irrelevant

Description of the problem including expected versus actual behavior:

The official documentation for path hierarchy tokenizer lacks usecase descriptions, as well as examples on performing filtering or aggregations on such fields. There is no explanation on why would anyone use this tokenizer.

Expected: the doc should explain why would I be using the tokenizer, and how do I aggregate/filter/etc. for the fields that use it.

Steps to reproduce: Doc issue, so no steps here.

clintongormley commented 8 years ago

It can certainly be improved. In the meantime, have a look at this: https://www.elastic.co/guide/en/elasticsearch/guide/master/denormalization-concurrency.html

s1monw commented 8 years ago

@kix may I ask you to reword the title and description of this issue and use less aggressive words. I'd appreciate if you could use terminology like Path hierarchy tokenizer doc is not helpful, needs to be improved, misses usecase. We are trying hard to get docs out there that are helpful as much as we try to continuously improve communication, please help me! :)

kix commented 8 years ago

@s1monw, done :)

s1monw commented 8 years ago

@kix thanks man!

kix commented 8 years ago

@s1monw, sorry for being a bit harsh, but sometimes when you're looking for a solution, you certainly hope that the doc will help.

s1monw commented 8 years ago

@kix not a big deal description looks much better and I can tell you there is suddenly more motivation to fix that :)

elasticmachine commented 6 years ago

Pinging @elastic/es-search-aggs

debadair commented 5 years ago

Need to pull relevant information into the ref from the Definitive Guide.

mayya-sharipova commented 5 years ago

Closing as this is addressed

mitoihs commented 2 years ago

I think it might be improved further. There's only an example of how to use that field to filter. I would be interested in:

  1. Is it possible to tokenize path hierarchy into a "keyword" type field? The example shows the "text" field but I don't want to run full-text searches on my tokens.
  2. How to aggregate on such a field? I would like to e.g. create a Pie Chart / Treemap that uses hierarchy. So it could look like a WinDirStat treemap:
  3. How to use a specific level in the hierarchy to filter/aggregate? (e.g.: I have a known number of levels each document will have and just want to filter on level 2)