Help with creating an OR query

[BenA-SA]

Hi all,

Brand new to opensearch and think this package is looking really helpful to me - thank you to those that made and maintain it.

One thing I'm really struggling with though is an OR query. I've got a list of 2-3 strings, and I want to return the documents where any of those strings is in 2 of the fields.

What I've been trying:

should_clauses = []

for string in strings:
    should_clauses.extend([
        Q("wildcard", field_1={"value": f"{string}"}),
        Q("wildcard", field_2={"value": f"*{string}*"}),
    ])
query = Q("bool", should=should_clauses, minimum_should_match="1")
search = CompanyDocument.search().query(query)

However I keep running into TypeErrors - unhashable type Bool (or Wildcard when I mess around with things).

I seem to get the output I'm looking for if I initialise it from a dict like this:

should_clauses = []
for string in strings:
    should_clauses.extend([
        {"wildcard" :{"field_1":{"value": f"*{string}*"}}},
        {"wildcard" :{"field_2":{"value": f"*{string}"}}},
    ])

query_dict = {
  "query": {
    "bool": {
      "should": should_clauses,
      "minimum_should_match": 1
    }
  }
}

s = CompanyDocument.search().from_dict(query_dict).execute()

Can anyone point me in the right direction?

Thanks for any help anyone can offer!

[qcoumes]

Your query seems valid, are you sure you're using django-opensearch-dsl's Qand not Django's ?

Can you give the result of pprint.pp(Q("bool", should=should_clauses, minimum_should_match="1").to_dict()) ?

[BenA-SA]

Thank you for taking the time to respond to me!

No problem - it's:

{'bool': {'should': [{'wildcard': {'field_1': {'value': '1234'}}},
                     {'wildcard': {'field_2': {'value': '*1234*'}}},
                     {'wildcard': {'field_1': {'value': '23456'}}},
                     {'wildcard': {'field_2': {'value': '*23456*'}}},
                     {'wildcard': {'field_1': {'value': '34567'}}},
                     {'wildcard': {'field_2': {'value': '*34567*'}}},
                     {'wildcard': {'field_1': {'value': '4567'}}},
                     {'wildcard': {'field_2': {'value': '*4567*'}}}],
          'minimum_should_match': '1'}}

weirdly, if I add to_dict() to my query (as below), it seems to work.

should_clauses = []

for string in strings:
    should_clauses.extend([
        Q("wildcard", field_1={"value": f"{string}"}),
        Q("wildcard", field_2={"value": f"*{string}*"}),
    ])
query = Q("bool", should=should_clauses, minimum_should_match="1")
search = CompanyDocument.search().query(query.to_dict())

I'm using opensearch-dsl 2.1.0 and django-opensearch-dsl==0.7.0

[qcoumes]

I can't reproduce your problem:

In [4]: from opensearchpy import Q

In [5]: should_clauses = []
    ...: strings = ["a", "b", "c"]
    ...: for string in strings:
    ...:     should_clauses.extend([
    ...:         Q("wildcard", field_1={"value": f"{string}"}),
    ...:         Q("wildcard", field_2={"value": f"*{string}*"}),
    ...:     ])
    ...: query = Q("bool", should=should_clauses, minimum_should_match="1")

In [6]: StayChild.search().query(query)
Out[6]: <django_opensearch_dsl.search.Search at 0x1264d7490>

It's weird that it's working after calling .to_dict() since it's basically what opensearchpy does.

I'm using django-opensearch-dsl 0.7.0 and opensearch-py 2.4.2 (we're not using opensearch-dsl anymore).

[BenA-SA]

No problem - at the moment it's only a proof of concept I'm testing, and it works fine with to_dict(), so that's fine for now. Thank you taking the time to try and assist!