Elasticsearch "starts with" search

Questions
,
1

Hi

I would like to include hits where the search string matches the start of a product name or article number.

For example a search for “stow” I would want to result in a hit for “stowaway”. Can that be solved by adding a wildcard(*?) or another analyzer?

I have tried something like this with no luck:

`allQueries.Add((qc.Match(x => x.Field(z => z.Name).Query(searchQuery.Text + "*").Boost(10).SynonymAnalyzer())
                                        || qc.Match(x => x.Field(z => z.ArticleNumber).Query(searchQuery.Text.ToLower()).Boost(2).SynonymAnalyzer())
                                        || qc.Match(x => x.Field(z => z.Content).Query(searchQuery.Text).Fuzziness(fuzziness).SynonymAnalyzer())));`

Litium version: 7.5.0

2

I had to do this on a project back in Litium 7.2, don’t know if it’s a better way:
|| qc.MatchPhrasePrefix(x => x.Field(z => z.ModelNames).Query(searchQuery.Text).Boost(1).MaxExpansions(100))

1 Like
3

Its working for Name, but when including ArticleNumber also I get an error from the Elasticsearch instance:

The remote server returned an error: (400) Bad Request.. Call: Status code 400 from: POST /prjproductdocument.sv-se/_search?typed_keys=true. ServerError: Type: search_phase_execution_exception Reason: "all shards failed"

Also I wonder if this solution works with synonyms?

4

For me, article number did not work at all, so i replaced it with a new field in the index and added this:
|| qc.MatchPhrasePrefix(x => x.Field(z => z.SearchableArticleNumber)

1 Like
5

The searching is dependent on the types of the fields that exists in the model. If we have the default ProductDocument-model from Accelerator we can see that the ArticleNumber property is decorated with the Keyword-attribute and that it not is supposed to use other than direct hits.

From Find strings within strings faster with the Elasticsearch wildcard field | Elastic Blog

when it came to searching strings, Elasticsearch has broadly offered two choices:

  1. Text fields
  2. Keyword fields

Text fields “tokenize” strings into multiple tokens, each typically representing words. Searchers looking for quick foxes can therefore match articles talking about a quick brown fox .

Keyword fields tend to be used for shorter, structured content like a country code which would be indexed without analysis as a single token. These can then be used by analysts e.g. to visualize popular holiday destinations using aggregations .

6

That works, thank you!

7

Ok, got it. Is it possible to boost hits on the first word in a tokenized text field? Or are they always weighted the same?

For example the customer would like searches for “stainless” to rank “Stainless pan” above “Sink stainless”

8

Maybe something as they suggest in this may help (I have not test this) Boosting score relative to location - Elastic Stack / Elasticsearch - Discuss the Elastic Stack but I don’t think it exists a one-property-change function to make it bosting like that.

9

I solved this by adding a property in the ProductDocument with the first word of the name, and included that property in the search

1 Like
10

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.