Indexing and searching a data collection

摘要

A system searches a data collection using a bloom filter index to reduce the search scope to improve search efficiency, while at the same time allowing more flexible searches of the data collection using search terms containing any arbitrary string of data.

主权项

1. A machine-implemented method as in claim of searching a collection of data, the method comprising:
generating a bloom filter for each item in a collection of data, the bloom filter having a set of elements, each element in the set of elements representing a fixed length run of characters contained in the item, the set of elements collectively representing every fixed length run of characters contained in the item, wherein generating the bloom filter for each item in a collection of data includes:
generating each fixed length run of characters contained in the item using a fixed length window positioned over the characters contained in the item,shifting the fixed length window over the characters contained in the item one character at time, andparsing the characters contained in the fixed length window at a given time into an element in the set of elements; receiving a request to search the collection of data for items containing a search term; filtering from the collection of data those items for which a fixed length run of characters contained in the search term is not represented in the set of elements of the respective items' bloom filters; and searching items not filtered from the collection of data for items containing the search term.