url提取

src/main/java/com/winhc/bigdata/udf/get_url.java

@@ -0,0 +1,37 @@
+package com.winhc.bigdata.udf;
+
+import com.aliyun.odps.udf.UDF;
+import org.apache.commons.lang.StringUtils;
+
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+/**
+ * @Author: π
+ * @Description: url 提取
+ */
+public class get_url extends UDF {
+
+    Pattern pattern2 = Pattern.compile("^((http://)|(https://))?([a-zA-Z0-9]([a-zA-Z0-9\\-]{0,61}[a-zA-Z0-9])?\\.)+[a-zA-Z0-9]{2,6}(/)");
+    Pattern pattern1 = Pattern.compile("http.+?html|http.+?cn|http.+?com|http.+/");
+
+    public String evaluate(String url) {
+        if (StringUtils.isBlank(url)) {
+            return "";
+        }
+        Matcher matcher = pattern2.matcher(url);
+        if (matcher.find()) {
+            System.out.println(matcher.group());
+            return matcher.group();
+        }
+        return "";
+    }
+
+    public static void main(String[] args) {
+        String str1 ="http://ssfw.zqdhfy.gov.cn/ktxx.aspx?cateId=15";
+        String str ="http://114.252.22.26.35/xx/yy/ktxx.aspx?cateId=15";
+        String res = new get_url().evaluate(str);
+        System.out.println(res);
+
+    }
+}