Author: Jim Teeuwen This package wraps the standard XML library and uses it to build a node tree of any document you load. This allows you to look up nodes forwards and backwards, as well as perform search queries (no xpath support). Nodes now simply become collections and don't require you to read them in the order in which the xml.Parser finds them. ================================================================================ DEPENDENCIES ================================================================================ goinstall go-charset.googlecode.com/hg/charset ================================================================================ USAGE ================================================================================ Getting the package up and running is simple enough. This should also automatically take care of any dependencies for you: $ goinstall github.com/jteeuwen/go-pkg-xmlx Using it: import xmlx "github.com/jteeuwen/go-pkg-xmlx" ================================================================================ API ================================================================================ The Document currently implements 2 simple search functions which allow you to look for specific nodes. *document.SelectNode(namespace, name string) *Node; *document.SelectNodes(namespace, name string) []*Node; SelectNode() returns the first, single node it finds matching the given name and namespace. SelectNodes() returns a slice containing all the matching nodes. Note that these search functions can be invoked on individual nodes as well. This allows you to search only a subset of the entire document. Each node exposes also a number of functions which allow easy access to a node value or an attribute value. They come in various forms to allow transparent conversion to types: int, int64, uint, uint64, float, float32, float64: *node.S(ns, name string) string *node.I(ns, name string) int *node.I64(ns, name string) int64 *node.U(ns, name string) uint *node.U64(ns, name string) uint64 *node.F(ns, name string) float *node.F32(ns, name string) float32 *node.F64(ns, name string) float64 *node.B(namespace, name string) bool Note that these functions actually consider child nodes for matching names as well as the current node. In effect they first perform a node.SelectNode() and then return the value of the resulting node converted to the appropriate type. This allows you to do this: Consider this piece of xml: red BMW Now this code: node := doc.SelectNode("", "car") brand := node.S("", "brand") Eventhough 'brand' is not the name of @node, we still get the right value back (BMW), because node.S() searches through the child nodes when looking for the value if the current node does not match the given namespace and name. For attributes, we only go through the attributes of the current node this function is invoked on: *node.As(ns, name string) string *node.Ai(ns, name string) int *node.Ai64(ns, name string) int64 *node.Au(ns, name string) uint *node.Au64(ns, name string) uint64 *node.Af(ns, name string) float *node.Af32(ns, name string) float32 *node.Af64(ns, name string) float64 *node.Ab(namespace, name string) bool All of these functions return either "" or 0 when the specified node or attribute could not be found. No errors are generated. The namespace name specified in the functions above must either match the namespace you expect a node/attr to have, or you can specify a wildcard "*". This makes node searches easier in case you do not care what namespace name there is or if there is one at all.