Chapter 5. Basic multifield search
This chapter covers
- Satisfying multiple user goals when searching
- Searching more than one field in your documents to satisfy user searches
- Transforming fields derived from your source data into a search-friendly form
- Balancing the influence of different search fields against one another
- Understanding the pros and cons of multifield search strategies
Earlier we compared search to a book’s index. Such an index lets you zero in on pages that discuss a subject you’re interested in. If you’re interested in the French Revolution, just browse to the back of your French history book and find the associated pages.
Similarly, a search engine can quickly identify documents that mention search terms by using an inverted index. Search for the term “revolution,” and the search engine retrieves a list of documents that mention a revolution. In the previous chapter, your goal was to use analysis to optimize the terms in the inverted index and maximize precision and recall. You expressed features of content as tokens, going beyond the idea that tokens are always associated with words and instead associating tokens with the meaning contained in the documents.