Question

In Elastic search, how can we search on free text. for example, we have free text data in one of the attribute like “test!£!£”!£”test“.

when, i do search using QueryBuilders, it gives exception due to special characters.

On searching web, come to know that we need to use escapeUtils from Lucene. QueryParser.escape(searchText)

Now, when i do exact match search in this case it ignores those characters and brings up the result for “test” as well.

Please suggest something on this.

Answer 1

You’re probably using the default mapping, thus the StandardAnalyzer. I guess it’s stripping out those characters, that are not even indexed. You’d need to decide how you want to index your text. Do you want to tokenize or do you want to have a single token? Do you want to keep those characters in the index? Based on that you need to pick the right analyzer and update your mapping through the Put mapping API. If you for example want to index the text as it is, you just configure the field as "index":"not_analyzed", which means that its context will be indexed but no text analysis will be applied to it.