A query-friendly compression for gml documents

Abstract

Geography Markup Language (GML) has become a standard encoding format for exchanging geographic data among heterogeneous Geographic Information System (GIS) applications. Whereas, the iteration of document structure and the textual expression of geographic data often cause the huge size of GML documents. In this paper, a query-friendly GML compression method is proposed, where the GML documents in SAX document parsing are transformed to a compact representation encompassing an event dictionary, the events hierarchy in balanced parentheses, a binary event wavelet tree and the document content blocks before compressed using a general compression utility. The proposed compression method supports direct path queries and spatial queries over the compressed files without the requirement of a full decompression. The compression model, the query resolution process and the compression algorithm are detailed in this paper, though the presentation is a preliminary investigation and it remains to carry out experiments to validate the proposed compression method on real GML documents.