How to include minimum_should_match in the query of Elasticsearch dsl?

Question

I’m trying to use the minimum_should_match how are you in documentation

q = Q('bool',
    must=[Q('match', title='python')],
    should=[Q(...), Q(...)],
    minimum_should_match=1
)
s = Search().query(q)

But when I include it in my query it returns nothing.

termo = 'Cartucho p/caneta tinteiro preto 301218 Pelikan CX 6 UN'

q = Q('bool', must=[Q('match', TermoBusca=termo)],minimum_should_match=1)
s = Search(using=es, index='produtos').query(q)
print(s.to_dict())

produtosEncontrados: list = s.execute()
for produto in produtosEncontrados:
    print(produto.Descricao)

If I remove the parameter (minimum_should_match=1) query returns results normally.

    termo = 'Cartucho p/caneta tinteiro preto 301218 Pelikan CX 6 UN'

    q = Q('bool', must=[Q('match', TermoBusca=termo)])
    s = Search(using=es, index='produtos').query(q)
    print(s.to_dict())

    produtosEncontrados: list = s.execute()
    for produto in produtosEncontrados:
        print(produto.Descricao)

Exit:

Generated query:

{'query': {'match': {'TermoBusca': 'Cartucho p/caneta tinteiro preto 301218 Pelikan CX 6 UN'}}}

Exit:

• Cartridge p/pen ink cartridge black 301218 Pelikan CX 6 UN
• Cartridge p/pen royal blue 301176 Pelikan CX 6 UN
• Ink p/pen 30ml black 301051 Pelikan EN 1 UN
• Ink p/pen 30ml dark blue 301028 Pelikan EN 1 UN
• Cartridge p/pen blue cartridge CA32005A Crown BT 3 UN
• Cartridge p/pen black cartridge CA32005P Crown BT 3 UN
• Capricci silver YW32616S Crown CX 1 UN source pen
• Capricci black ink pen YW32616P Crown CX 1 UN
• Neutral Regent Source Pen YW39422N Crown CX 1 UN
• Cartridge p/Brother black LC3039BK Brother CX 1 UN

I’ve looked everywhere and I don’t understand what I’m doing wrong. My intention is to filter through the field TermoBusca 100% of the term reported. As in the query below, only using the DSL.

    termo = 'Cartucho p/caneta tinteiro preto 301218 Pelikan CX 6 UN'
    query_body = {
            "sort" : [
                { "Prioridade" : {"order" : "desc"}},
                { "CurvaABC" : {"order" : "desc"}},
                "_score"
            ],
            "query": {
                "match": {
                    "TermoBusca": {
                        "query": termo,
                        "minimum_should_match": "100%"
                    }
                }
            },
        "size": 60
    }

    result = es.search(index="produtos", body=query_body)
    print()

    for hit in result['hits']['hits']:
        print(hit['_source']['TermoBusca'])
    self.assertEqual(56, len(result['hits']['hits']))

Exit:

• Cartridge p/pen ink cartridge black 301218 Pelikan CX 6 UN

Answer 1

The parameter minimum_should_match should be used in conjunction with optional clauses of the type should. When minimum_should_match=1 an item will only be returned by the query if at least one of the clauses should is true. Your query has no clauses should. She has only one clause must. By definition, when you use must, all returned items must satisfy the clause condition. Hence, your query with must already filtered by the field TermoBusca in 100% of cases (the use of minimum_should_match is not necessary).

In older versions of Elasticsearch (before version 5), in the absence of clauses should the parameter minimum_should_match was ignored. This behavior has changed and in modern versions of Elasticsearch. Nowadays a query with minimum_should_match greater than the number of clauses should does not return any item.

This change impacted some systems that built clauses should dynamically.

I had to deal with this problem in the past after an update of Elasticsearch broke a system’s search functionality. In this case, when no optional filter was specified the searches stopped returning results. The solution was to modify the logic that built the search to only include the parameter minimum_should_match when the amount of clauses should is greater than 1. More details about this problem can be found in the following Issue of Github (in English):

• Issue #21844: Bool 'must' query with "minimum_should_match": "1", then result is Empty

---

Update:

The mass of data posted by Marconcilio called my attention to the fact that the parameter minimum_should_match is in the query internal. The desired behavior of the OP, in this case, is that the query only returns items where TermoBusca contains 100% of the terms in the string termo. In the DSL of Python we can rebuild the internal query with a dict:

 q = Q('match', TermoBusca={'query': termo, 'minimum_should_match': '100%'})