Merge pull request #12 from TuSKan/master

#11 [New feature]: ParseAttributesOnly
tamerh · Mar 2, 2020 · c5f7629 · c5f7629
2 parents ac36c34 + 7a15c57
commit c5f7629
Show file tree

Hide file tree

Showing 4 changed files with 52 additions and 11 deletions.
diff --git a/README.md b/README.md
@@ -6,7 +6,7 @@ xml-stream-parser is xml parser for GO. It is efficient to parse large xml data
 
 ```xml
 <?xml version="1.0" encoding="UTF-8"?>
-<bookstore>
+<bookstore number="2" loc="273456">
    <book>
       <title>The Iliad and The Odyssey</title>
       <price>12.95</price>
@@ -52,6 +52,12 @@ for xml := range parser.Stream() {
 parser := xmlparser.NewXMLParser(br, "book").SkipElements([]string{"price", "comments"})
 ```
 
+**Attributes** only
+
+```go
+parser := xmlparser.NewXMLParser(br, "bookstore", "book").ParseAttributesOnly("bookstore")
+```
+
 **Error** handlings
 
 ```go

diff --git a/sample.xml b/sample.xml
@@ -7,9 +7,9 @@
 <!ELEMENT para (#PCDATA)>
 ]>
 <!-- Comment at top after declarations -->
-<examples>
+<examples att1="ex1" att2="ex2">
     <tag1 att1="<att0>" att2="att0">
-        <tag11 att1="att0">Hello                                    <![CDATA[你好]]>            <!-- comment --><![CDATA[Gür]]></tag11>
+        <tag11 att1="att0">Hello                                                <![CDATA[你好]]>            <!-- comment --><![CDATA[Gür]]></tag11>
         <tag11 att1="att0">InnerText111</tag11>
         <tag12 att1="att0"/>
         <tag13>InnerText13</tag13>

diff --git a/xmlparser.go b/xmlparser.go
@@ -12,6 +12,7 @@ type XMLParser struct {
 	loopElements      map[string]bool
 	resultChannel     chan *XMLElement
 	skipElements      map[string]bool
+	attrOnlyElements  map[string]bool
 	skipOuterElements bool
 	xpathEnabled      bool
 	scratch           *scratch
@@ -22,12 +23,13 @@ type XMLParser struct {
 func NewXMLParser(reader *bufio.Reader, loopElements ...string) *XMLParser {
 
 	x := &XMLParser{
-		reader:        reader,
-		loopElements:  map[string]bool{},
-		resultChannel: make(chan *XMLElement, 256),
-		skipElements:  map[string]bool{},
-		scratch:       &scratch{data: make([]byte, 1024)},
-		scratch2:      &scratch{data: make([]byte, 1024)},
+		reader:           reader,
+		loopElements:     map[string]bool{},
+		attrOnlyElements: map[string]bool{},
+		resultChannel:    make(chan *XMLElement, 256),
+		skipElements:     map[string]bool{},
+		scratch:          &scratch{data: make([]byte, 1024)},
+		scratch2:         &scratch{data: make([]byte, 1024)},
 	}
 
 	// Register loop elements
@@ -49,6 +51,13 @@ func (x *XMLParser) SkipElements(skipElements []string) *XMLParser {
 
 }
 
+func (x *XMLParser) ParseAttributesOnly(loopElements ...string) *XMLParser {
+	for _, e := range loopElements {
+		x.attrOnlyElements[e] = true
+	}
+	return x
+}
+
 // by default skip elements works for stream elements childs
 // if this method called parser skip also outer elements
 func (x *XMLParser) SkipOuterElements() *XMLParser {
@@ -136,7 +145,9 @@ func (x *XMLParser) parse() {
 					continue
 				}
 
-				element = x.getElementTree(element)
+				if _, ok := x.attrOnlyElements[element.Name]; !ok {
+					element = x.getElementTree(element)
+				}
 				x.resultChannel <- element
 				if element.Err != nil {
 					return

diff --git a/xmlparser_test.go b/xmlparser_test.go
@@ -50,7 +50,7 @@ func TestBasics(t *testing.T) {
 		panic("Test failed")
 	}
 
-	if results[0].Childs["tag11"][0].InnerText != "Hello                                    你好            Gür" {
+	if results[0].Childs["tag11"][0].InnerText != "Hello                                                你好            Gür" {
 		panic("Test failed")
 	}
 
@@ -408,6 +408,30 @@ func TestXpathNS(t *testing.T) {
 
 }
 
+func TestAttrOnly(t *testing.T) {
+	p := getparser("examples", "tag1").ParseAttributesOnly("examples")
+	for xml := range p.Stream() {
+		if xml.Err != nil {
+			t.Fatal(xml.Err)
+		}
+		if xml.Name == "examples" {
+			if len(xml.Childs) != 0 {
+				t.Fatal("Childs not empty for ParseAttributesOnly tags")
+			}
+			fmt.Printf("Name: \t%s\n", xml.Name)
+			fmt.Printf("Attrs: \t%v\n\n", xml.Attrs)
+		}
+		if xml.Name == "tag1" {
+			if len(xml.Childs) == 0 {
+				t.Fatal("Childs not empty for ParseAttributesOnly tags")
+			}
+			fmt.Printf("Name: \t%s\n", xml.Name)
+			fmt.Printf("Attrs: \t%v\n", xml.Attrs)
+			fmt.Printf("Childs: %v\n", xml.Childs)
+		}
+	}
+}
+
 func Benchmark1(b *testing.B) {
 
 	for n := 0; n < b.N; n++ {