I am using LSH for database records and by that I am creating a index (not a database index, a simple hashmap) where similar records blocked in to the same bucket. The database may contain several millions of records. My question regards with the design I post below.
First I will create the index using the database available by executing LSH. But when a new record inserted in to the database I must index that record also in to the index. How can I do this using LSH? Can LSH allocate that record to the bucket that have similar records?? Does LSH support updates in to the dataset?