Difference Computation for Grammar-Compressed XML Data

Whenever web data processing requires the storage or the exchange of multiple versions of big XML data collections and the pure size of big XML data becomes a bottleneck in storage or fast data exchange over the web, XML compression and XML version control may become significant contributions to avoid such a bottleneck. Grammar-based compression is of increasing importance for big XML data collections in the web as it allows fast queries and updates on compressed data without full decompression. However, merging different versions of grammar-based compressed XML data collections is a challenge, because small differences in two given uncompressed XML files may lead to significant differences in the grammar-based compressed data formats of these files. Therefore, when multiple versions of an XML file have to be stored in compressed form, the different compressed formats may be difficult to combine, which weakens the benefit achieved by the compression. To overcome this weakening, we present a technique to compute the common part and the difference of two compressed XML documents without the need to fully decompress the documents. Our approach computes a compressed common prefix and parameters representing the difference of two compressed XML documents in polynomial time in the size of the grammar compressed documents, even if the common part of the documents is hidden in completely different sets of compressed grammar rules.