Matching Exact Values with ArangoSearch

Search for exact strings, numbers or number ranges, as well as booleans without Analyzer transformations applied

If you want to find strictly equal values, then the identity Analyzer is what you need. It will not apply any transformations. It is a no-operation Analyzer that passes everything through unaltered.

Matching an Exact String

You can index and search strings with the identity Analyzer for exact matching, that is case-sensitive, with accented characters as-is, and only if the entire string is equal (not matching substrings).

Dataset

IMDB movie dataset

View definition

search-alias View

db.imdb_vertices.ensureIndex({ name: "inv-exact", type: "inverted", fields: [ "title" ] });
db._createView("imdb", "search-alias", { indexes: [ { collection: "imdb_vertices", index: "inv-exact" } ] });

arangosearch View

{
  "links": {
    "imdb_vertices": {
      "fields": {
        "title": {
          "analyzers": [
            "identity"
          ]
        }
      }
    }
  }
}

AQL queries

Match exact movie title (case-sensitive, full string):

FOR doc IN imdb
  SEARCH doc.title == "The Matrix"
  RETURN doc.title

For arangosearch Views, it is not necessary to set the Analyzer context with the ANALYZER() function here, because the default Analyzer is identity anyway. However, being explicit about the Analyzer context can help with clarity and it also makes it easier to adjust queries if you want to use something other than the identity Analyzers:

FOR doc IN imdb
  SEARCH ANALYZER(doc.title == "The Matrix", "identity")
  RETURN doc.title

A common pitfall is to index a field with a certain Analyzer, but forgetting to set this Analyzer as context in the query. The consequence is usually an empty result, because there is nothing in the View index for the implicitly requested Analyzer. Or the field happens to be indexed with the identity Analyzer as well, but there are no or different matches because of different Analyzer pre-processing between the indexed data and the search terms.

For search-alias Views, you never need to specify an Analyzer in queries, because the Analyzer is inferred from the inverted index definition, which only allows a single Analyzer per field.

Matching Multiple Strings

You can search for several alternatives by combining multiple conditions with logical ORs, but you can also use the IN operator for the same purpose. The advantage of the IN operator is that it scales better with the number of alternatives and allows you to use a single query for a varying amount of strings that you want to match.

Match multiple exact movie titles using OR:

FOR doc IN imdb
  SEARCH doc.title == "The Matrix" OR doc.title == "The Matrix Reloaded"
  RETURN doc.title

Match multiple exact movie titles using IN:

FOR doc IN imdb
  SEARCH doc.title IN ["The Matrix", "The Matrix Reloaded"]
  RETURN doc.title

By substituting the array of strings with a bind parameter, it becomes possible to use the same query for an arbitrary amount of alternative strings to match:

FOR doc IN imdb
  SEARCH doc.title IN @titles
  RETURN doc.title

Bind parameters:

{
  "titles": [
    "The Matrix",
    "The Matrix Reloaded",
    "The Matrix Revolutions",
    "The Matrix Trilogy",
    "The Matrix Revisited"
  ]
}

Using Negations

Searching for exact values does not end with one or many equality conditions, but extends to negations as well. You can search for inequality with the != operator to return everything from the View index but documents that do not fulfill the criterion. This is also works with multiple alternatives using the NOT IN operator.

Match movies that do not have the title The Matrix:

FOR doc IN imdb
  SEARCH doc.title != "The Matrix"
  RETURN doc.title

Note that this includes documents that do not even have a title attribute, with the effect of returning many null values in the result.

Match movies that neither have the title The Matrix nor The Matrix Reloaded. Post-filter the results to exclude implicit nulls:

FOR doc IN imdb
  SEARCH doc.title NOT IN ["The Matrix", "The Matrix Reloaded"]
  FILTER doc.title != null
  RETURN doc.title

A better way to ignore documents without title attribute is to use the EXISTS() function. For arangosearch Views, you need to change the storeValues View property (not to be confused with storedValues!) from "none" to "id". You can then to test whether there is a title field or not. For search-alias Views, the EXISTS() function works out of the box.

On a single server with this particular dataset, the query is roughly five times faster than the previous one without EXISTS():

FOR doc IN imdb
  SEARCH EXISTS(doc.title) AND doc.title NOT IN ["The Matrix", "The Matrix Reloaded"]
  RETURN doc.title