Documentation wrong: still describes index=analyzed/not_analyzed for type string #30122
Assignees
Labels
>docs
General docs changes
:Search Relevance/Analysis
How text is split into tokens
Team:Search Relevance
Meta label for the Search Relevance team in Elasticsearch
Comments
|
Pinging @elastic/es-search-aggs |
jpountz
added a commit
that referenced
this issue
Apr 25, 2018
jpountz
added a commit
that referenced
this issue
Apr 25, 2018
jpountz
added a commit
that referenced
this issue
Apr 25, 2018
|
This was fixed in the 6.4 docs. The original issue refers to the Elasticsearch Definitive Guide, which has not been updated past version 2.x |
javanna
added
the
Team:Search Relevance
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Elasticsearch version: 6.2
Plugins installed: []
JVM version (
java -version): 8OS version (
uname -aif on a Unix-like system): Windows Server 2016 (or whatever you like)Description of the problem including expected versus actual behavior:
Documentation still explains to model type "string" with a property "index" set to "analyzed", "not_analyzed" or "no", whereas it appears that neither is valid anymore since version 5.0
From what I read in the wild wild web it should be type=text (for analyzed) or type=keyword (for not_analyzed), and i guess "no" has become index=false.
Thus the documentation is currently plain wrong, and pushes the interested user into the abyss of google.
Steps to reproduce:
Wrong explanation can be found at least here: https://www.elastic.co/guide/en/elasticsearch/guide/current/mapping-intro.html
Other places are of course possible.
Provide logs (if relevant):
The text was updated successfully, but these errors were encountered: