Indexing Text Documents

• Inverted index

  • Maps each word to a list of docs that contain it

• For fast search through inverted index

  • B+ tree or hash index (lexicon) on all terms

• Boolean query: intersect/merge doc lists

• Ranked query:

  • Merge doc lists and for each document
  • Compute relevance with respect to query
  • Fetch and return docs in decreasing rank order

Notes: