Compressed Data Structures for Astronomical Content-Aware Resource Search

We introduce an efficient approach that aims at supporting content-based queries on the Chilean Virtual Observatory. In particular we are interested in retrieving relevant information from virtual-observatory tables. This introduces several challenges that make the information-retrieval process harder. We define an algorithm that uses a compressed data structure to obtain the count of the number of occurrences of a string query within each column of a table. This kind of query has been used in the literature for faceted and semantic search as well as for retrieving information from web tables. This is in order to improve search effectiveness. We show that using only 15%-25% the space of a table our approach contains the table data (and hence the table can be deleted) and is able to answer queries efficiently in a few milliseconds.