CODEWITHSUNDEEP
elasticsearch
Antonio Romero Oca
Antonio Romero Oca

Reputation: 1245

Problems with large bool on elasticsearch

I need to get all documents that contain at least one item from a list of more than 1024.

My query is basically a bool query with a should and minimum_should_match: 1.

Elasticsearch maxClauseCount is set to 1024 by default. I have tried to set it to 4096 and the configuration looks to be ok:

I request http://myserver:9200/my_index/_settings and get:

... "query": { "bool": { "max_clause_count": "4096" } } ..., But if I try still get TooManyClauses[maxClauseCount is set to 1024] in my logs.

1st Question: Why is this contradictory?

I have read that for some cases it is better to use a filter instead of a large bool:

In general I'd recommend to rewrite that query to use terms filter instead of boolean query https://discuss.elastic.co/t/too-many-clauses-maxclausecount-is-set-to-1024/61968

2nd Question: How could I use a filter to obtain the same logic as the multiple should bool on my example? What is best bool filter or filtered filter for that case?

Upvotes: 2

Views: 1367

Answers (1)

Antonio Romero Oca
Antonio Romero Oca

Reputation: 1245

I am not yet sure why Elasticsearch raise the maxClauseCount error, but i have found an alternative way to structure my query.

The (easy) solution is to use terms with the large list of items. If I use it into a must I get the same error, but with filter it works perfectly.

Example: 

{
  "query": {
    "bool": {
      "filter": [
        {"terms": {"my_field": ["item1", "item2", ... "itemN"]}}
      ]
    }
  }
}

The only lack for filter is that:

The clause (query) must appear in matching documents. However unlike must the score of the query will be ignored. https://www.elastic.co/guide/en/elasticsearch/reference/2.3/query-dsl-bool-query.html

Upvotes: 2

Related Questions