The projects aims to be xml deserializer working somewhere between SAX and DOM model. It benefits from multiple CPUs and fast SDD. It is multiple times faster then built-in .net XmlReader.
Just include latest NuGet package.
using (FileStream stream = File.OpenRead(path))
{
HashSet<string> found = new HashSet<string>();
XenonReader reader = new XenonReader(source, "book");
await reader.Process(document =>
{
found.Add(document.Get("id").ToString());
});
Console.WriteLine($"Found {found.Count} unique books.");
}-
Intel(R) Core(TM)2 Quad CPUQ9300 @ 2.50GHz, 6GB RAM, TOSHIBA MD04ACA50D (max 200MB/s)
-
Source: plwiki-20170801-pages-meta-current4.xml (about 5GB)
Test Case Xenon Static Xenon Dynamic Xml Reader just-seek 31.6s, 22.1s CPU 31.6s, 21.6s CPU 60.3s, 60.2s CPU text-length 31.6s, 41.4s CPU 31.6s, 42.1s CPU 61.8s, 61.1s CPU -
Intel(R) Core(TM) i7-4770HQ CPU @ 2.20GHz, 16GB RAM, APPLE SSD SM0512G (max 1500MB/s)
-
Source: plwiki-20170901-pages-meta-history2.xml (about 100GB)
Test Case Xenon Static Xenon Dynamic Xml Reader just-seek 76.7s, 183.5s CPU 75.1s, 181.5s CPU 570.0s, 574.8s CPU text-length 94.2s, 641.9s CPU 94.1s, 638.0s CPU 595.3s, 594.5s CPU