MongoDB: Text index with variable weights

Question

The MongoDB documentation lists this as an example command to add a text index on fields using weights:

db.blog.ensureIndex(
                     {
                       content: "text",
                       keywords: "text",
                       about: "text"
                     },
                     {
                       weights: {
                                  content: 10,
                                  keywords: 5,
                                },
                       name: "TextIndex"
                     }
                   )

If I have another field, say 'weight_value', in each document with a varying number, could I use this field value instead of a discrete number?

Thanks,

score 1 · Answer 1 · answered Feb 21 '14 at 10:13

Since you are quoting from the manual page a general perusal of the rest of the content there should confirm for you that text searching in MongoDB does not work that way. So in addition to the section you have quoted consider the following document:

{ _id: 2,
  content: "Who doesn't like cake?",
  about: "food",
  keywords: [ "cake", "food", "dessert" ]
}

From the information have shown above, in the context of this document the weighting for any search terms will be ordered by

Highest weighting to matches in content 10
Next highest weighting to matches in keywords 5
Default weighting to the indexed field about 1

So placing your own "weighting factor" in the document will not work as it may in some other searching solutions. Such as giving a "boost" to certain documents regardless of the relevance match.

You could work within the constraints of the provided mechanisms, or if this does not absolutely suit you, then perhaps look at offloading your text search capabilities for you application to one of the more specialized services.

MongoDB: Text index with variable weights

1 Answers1