|sanatizers||Misspelled Keyword giving null results.|
It would be good if you put a bit more effort into describing the issue and show the query.
wrong keyword spelling - sanatizers,
correct keyword spelling - sanitizers,
//'dfs_query_then_fetch' => true,
|Misspelled Keyword giving null results.|
That is not what lenient does. I think you need to specify fuzziness in the query to catch that case.
I am not saying about lenient,
I want to know that,
IS elastic search checks wrong keywords spelling?
Keyword fields are by default exact match only, which is why you need to use fuzziness to get the behaviour you are describing.
Can you give any suggestions and examples for fuzziness?
I do not have any examples. I would recommend looking at the ones in the docs.
Again, please don't post duplicate posts.
I am using stemmer for singular and plural keywords, is it better?
The honest answer is "we don't know".
It depends on your data.
It depends on your queries.
It depends if you want to optimise for relevance or selling products with high markup.
It depends if you favour precision.
It depends if you favour recall.
It depends if you want fast search.
It depends if you want low compute cost.
A better first question to ask yourself is this:
"How will I know if [insert configuration choice here] is beneficial to my site?"
You can't improve what you can't measure. Figure out how you will measure "better" first then you will have the tools to make better configuration choices with more certainty.
"Relevance engineer" for an ecommerce site is a skilled job and it starts with bringing some science to how you measure the effects of changes. It's also an ongoing process rather than a one-time config setup. You will get the answers you need not from here but from your users, your data and the systems you put in place to monitor how well you're connecting those people to that data.
You asked the same question in Customer Interest Related Keyword search in Elastic Search. Let's keep the discussion in one single place. Thanks.
@Mark_Harwood: What insert configuration is better for ecommerce products search of all types?
What's the best car?
As I said, let's keep the discussion in Customer Interest Related Keyword search in Elastic Search
I'm closing this thread.