本文内容由阿里云实名注册用户自发贡献，版权归原作者所有，阿里云开发者社区不拥有其著作权，亦不承担相应法律责任。具体规则请查看《
阿里云开发者社区用户服务协议
》和
《
阿里云开发者社区知识产权保护指引
》。如果您发现本社区中有涉嫌抄袭的内容，填写
侵权投诉表单
进行举报，一经查实，本社区将立刻删除涉嫌侵权内容。
在这里对页面各字段进行一下说明，从上往下：
1. Apply to，是指你提取的变量要应用的范围，需要注意的是，Jmeter Variable是指应用到全局，也就是跨线程的。
2. 要检查的响应字段，是指你的正则表达式的提取范围，主体应该是整个response，单选body就是指在body里找，其它选项根据字面意思，就不再赘述。
3. 引用名称，是我们自定义的一个变量名称，也就是提取出来的值的变量名称，例如我给它定义为“var”，那么正则提取出来的值就存在var里，之后如果要用，就使用el表达式，也就是${var（刚刚定义的名称）}的方式来使用。
4. 正则表达式，这里写提取变量用的
正则表达式
。
5. 模板，这里是指你提取多少个变量（如果在正则表达式中有多个提取表达式），一般我们提取一个就够了，模板采用$N$的方式，N代表数量，例如，我们设立一个只提取1个值的模板，就写作$1$。
6. 匹配数字，也就是提取匹配到的第几个值，例如写1就是从匹配到的值里选取第一个。
7. 缺省值，也就是如果提取不到，那么默认给一个值替换提取结果。
在这里举一个例子如下：
现在有一段response值，
{"errorCode":"","success":true,"message":"","data":{"scoreId":"98","templateId":"","title":"新建权重评分表-auto-test","type":"1","state":"2","containers":[{"id":"692","text":"维度1","components":[{"id":"2196","text":"第一个指标","percent":"25","type":"1","options":[{"id":"6978","text":"好","score":"100","sort_num":1},{"id":"6979","text":"较好","score":"80","sort_num":2},{"id":"6980","text":"一般","score":"60","sort_num":3},{"id":"6981","text":"较差","score":"40","sort_num":4},{"id":"6982","text":"差","score":"0","sort_num":5}]},{"id":"2197","text":"第二个指标","percent":"25","type":"2","options":[{"id":"6983","text":"评分标准1","score":"100","sort_num":1}]}]},{"id":"693","text":"维度2","components":[{"id":"2198","text":"第三个指标","percent":"25","type":"2","options":[{"id":"6984","text":"评分标准2","score":"100","sort_num":1}]},{"id":"2199","text":"第四个指标","percent":"25","type":"1","options":[{"id":"6985","text":"好","score":"100","sort_num":1},{"id":"6986","text":"较好","score":"80","sort_num":2},{"id":"6987","text":"一般","score":"60","sort_num":3},{"id":"6988","text":"较差","score":"40","sort_num":4},{"id":"6989","text":"差","score":"0","sort_num":5}]}]}]}}
我们需要从中把维度1对应的id提取出来，那么可以按照如下参数（."id":"(.+?)","text":"维度1"*）提取：
beanshell就是一个运行Java的微环境（大概可以这么理解），在Script框里我们写java代码，jmeter执行到这里的时候就会执行里面的java代码，beanshell里面会有一些jmeter内置的变量，用于jmeter和你写的Java代码交互，常用的大约有这么几个：
1. log：写入信息到jmeber.log文件，使用方法：log. info(“This is log info!”);
2. vars - (JMeterVariables)：操作jmeter变量，这个变量实际引用了JMeter线程中的局部变量容器（本质上是Map），它是测试用例与BeanShell交互的桥梁，常用方法：
vars.get(String key)：从jmeter中获得变量值
vars.put(String key，String value)：数据存到jmeter变量中
3. prev - (SampleResult)：获取前面的sample返回的信息，常用方法：
getResponseDataAsString()：获取响应信息
getResponseCode() ：获取响应code
在这里根据我在工作里遇到的一个实战来说明
首先说明下需求，现在存在一个接口是post类型，http协议的，他需要若干个参数用以传参，其中一个参数“scoreData”是json类型的数据，数据结构如下：
{
"645（题目的id）":{"id":"2546（选项的id）","percent":"25（分值，固定值）"},
"646（同上）":{"id":"2551（同上）","percent":"24.75（同上）","value（该参数只有在题目的type=2的时候才会出现）":"99（分值，固定值）"}
}
说明一下这个数据结构，我们的问卷可能会存在多个题目，上面case里的645，646就代表是两道题目，题目的类型分为选择题和填分题，选择题的type是1，我们只取第一个选项的id（也就是说传参代表勾选了第一项）；填分题的type是2，我们这里固定填99分，也就是value是99；percent这个字段是根据权重计算出来的值，这里不多做说明，就当做它的值也是一个固定的值。
数据全部来源于前一个接口的返回值，返回值如下：
{"success":true,"message":"","errorCode":"","data":[{"submit_state":"0","score_id":"f5","travel_id":"5","travel_title":"auto-test-zyj1502356486388","inspect_time":"2033-08-23","category_codes":"2266","suggest_content":"","my_score":"0","categoryName":"商业包装设计与施工","scoreData":{"fbScoreId":"f5","templateId":"","title":"新建权重评分表-auto-test-zyj","type":1,"containers":[{"id":"8","text":"维度1","components":[{"id":"14","text":"第一个指标","percent":"25","type":"1","options":[{"id":"38","text":"好","score":"100","sort_num":1},{"id":"39","text":"较好","score":"80","sort_num":2},{"id":"40","text":"一般","score":"60","sort_num":3},{"id":"41","text":"较差","score":"40","sort_num":4},{"id":"42","text":"差","score":"0","sort_num":5}],"select":""},{"id":"15","text":"第二个指标","percent":"25","type":"2","options":[{"id":"43","text":"评分标准1","score":"100","sort_num":1}],"select":""}]},{"id":"9","text":"维度2","components":[{"id":"16","text":"第三个指标","percent":"25","type":"2","options":[{"id":"44","text":"评分标准2","score":"100","sort_num":1}],"select":""},{"id":"17","text":"第四个指标","percent":"25","type":"1","options":[{"id":"45","text":"好","score":"100","sort_num":1},{"id":"46","text":"较好","score":"80","sort_num":2},{"id":"47","text":"一般","score":"60","sort_num":3},{"id":"48","text":"较差","score":"40","sort_num":4},{"id":"49","text":"差","score":"0","sort_num":5}],"select":""}]}]},"travelDetail":{"travel_id":"5","score_id":"f5","travel_title":"auto-test-zyj1502356486388","company_name":"上海东方雨虹防水技术有限责任公司","supplier_type":"2","tax_qualify":"2","organizer_id":"195104","inspect_time":"2033-08-23","scoreName":"新建权重评分表-auto-test-zyj","supplierTypeName":"代理商","taxQualifyName":"小规模纳税人","organizerName":"125","Teams":"zyj账号01、zyj账号02","Locations":[{"name":"办公室","remark":"备注1"},{"name":"地点一","remark":"备注2"}],"QualifyFiles":[{"name":"营业执照","remark":"备注3"},{"name":"地点二","remark":"备注4"}]},"travelPhoto":[],"photoDescription":{"1":[{"address":"地点一"},{"address":"办公室"}],"2":[{"address":"地点二"},{"address":"营业执照"}]},"photo_count":0,"score_item_total_count":"4"},{"submit_state":"0","score_id":"f4","travel_id":"4","travel_title":"auto-test-zyj1502351094326","inspect_time":"2033-08-23","category_codes":"2266","suggest_content":"","my_score":"0","categoryName":"商业包装设计与施工","scoreData":{"fbScoreId":"f4","templateId":"","title":"新建权重评分表-auto-test-zyj","type":1,"containers":[{"id":"6","text":"维度1","components":[{"id":"10","text":"第一个指标","percent":"25","type":"1","options":[{"id":"26","text":"好","score":"100","sort_num":1},{"id":"27","text":"较好","score":"80","sort_num":2},{"id":"28","text":"一般","score":"60","sort_num":3},{"id":"29","text":"较差","score":"40","sort_num":4},{"id":"30","text":"差","score":"0","sort_num":5}],"select":""},{"id":"11","text":"第二个指标","percent":"25","type":"2","options":[{"id":"31","text":"评分标准1","score":"100","sort_num":1}],"select":""}]},{"id":"7","text":"维度2","components":[{"id":"12","text":"第三个指标","percent":"25","type":"2","options":[{"id":"32","text":"评分标准2","score":"100","sort_num":1}],"select":""},{"id":"13","text":"第四个指标","percent":"25","type":"1","options":[{"id":"33","text":"好","score":"100","sort_num":1},{"id":"34","text":"较好","score":"80","sort_num":2},{"id":"35","text":"一般","score":"60","sort_num":3},{"id":"36","text":"较差","score":"40","sort_num":4},{"id":"37","text":"差","score":"0","sort_num":5}],"select":""}]}]},"travelDetail":{"travel_id":"4","score_id":"f4","travel_title":"auto-test-zyj1502351094326","company_name":"上海东方雨虹防水技术有限责任公司","supplier_type":"2","tax_qualify":"2","organizer_id":"195104","inspect_time":"2033-08-23","scoreName":"新建权重评分表-auto-test-zyj","supplierTypeName":"代理商","taxQualifyName":"小规模纳税人","organizerName":"125","Teams":"zyj账号01、zyj账号02","Locations":[{"name":"办公室","remark":"备注1"},{"name":"地点一","remark":"备注2"}],"QualifyFiles":[{"name":"营业执照","remark":"备注3"},{"name":"地点二","remark":"备注4"}]},"travelPhoto":[],"photoDescription":{"1":[{"address":"地点一"},{"address":"办公室"}],"2":[{"address":"地点二"},{"address":"营业执照"}]},"photo_count":0,"score_item_total_count":"4"}]}
要提取数据，尤其是这么大的json数据，首先我们要把这个json的数据结构分析清楚，我推荐大家把这个json放到json解析网站里解析一下结构，这样会更清楚，例如bejson，解析出来如下（部分要点截图）：
将取到的值组合成需要的数据结构再输出。
对于json的处理，我们需要用到额外的包，json的官网里有，在java选项下面，第一个JSON-java就是，这个包叫做org.json,我们下载下来之后，把这个包放到jmeter的apache-jmeter\lib\ext下面，这样就可以用了。
简单介绍下org.json，这部分建议看一下教程；
JSONObject
是一个无序的键/值对集合。
它的表现形式是一个包裹在花括号的字符串，键和值之间使用冒号隔开，键值和键值之间使用逗号隔开。
内在形式是一个使用get()和opt()方法通过键来访问值，和使用put()方法通过键来添加或者替代值的对象。
值可以是任何这些类型：Boolean,JSONArray,JSONObject,Number和String，或者JOSONObject.NULL对象。
JSONArray
是一个有序的序列值。
它的表现形式是一个包裹在方括号的字符串，值和值之间使用逗号隔开。
内在形式是一个使用get()和opt()方法通过索引来访问值，和使用put()方法来添加或修改值的对象。
值可以是任何这些类型：Boolean,JSONArray,JSONObject,Number,和String，或者JSONObject.NULL对象。
我们大约只用到这两种就可以了
这里直接放出处理的源码：
import java.util.HashMap;
import java.util.Map;
import org.json.*;
System.out.println("test................................................................");
//提取返回值作为变量response_data
String response_data = prev.getResponseDataAsString();
//解析json
JSONObject data_obj = new JSONObject(response_data);
JSONArray jsonarray = data_obj.getJSONArray("data");
JSONObject scoreData = null;
int id = Integer.parseInt(vars.get("travel_id"));
System.out.println("id="+id);
for (int i = 0; i < jsonarray.length(); i++) {
JSONObject jsonobj = jsonarray.getJSONObject(i);
System.out.println(jsonarray.optString(i));
System.out.println(jsonobj.getInt("travel_id"));
System.out.println(jsonobj.getString("travel_title"));
int travle_id = jsonobj.getInt("travel_id");
if(travle_id==id){
scoreData = jsonobj.getJSONObject("scoreData");
JSONArray containers = scoreData.getJSONArray("containers");
Map data = new HashMap();
for (int i = 0; i < containers.length(); i++){
JSONArray component_ary = containers.getJSONObject(i).getJSONArray("components");
for (int j = 0; j < component_ary.length(); j++){
System.out.println(component_ary.optString(j));
System.out.println(component_ary.getJSONObject(j).get("id"));
String data_id = component_ary.getJSONObject(j).get("id").toString();
JSONArray options = component_ary.getJSONObject(j).getJSONArray("options");
String score_id = options.getJSONObject(0).get("id").toString();
System.out.println(options.getJSONObject(0).get("id"));
Map score = new HashMap();
score.put("id", score_id);
score.put("percent", "25");
if(component_ary.getJSONObject(j).getInt("type")==2){
score.put("value", "99");
System.out.println(score);
data.put(data_id, score);
System.out.println(data);
//将map类型转换为json字符
JSONObject putdata = new JSONObject(data);
System.out.println(putdata);
//将putdata输出到jmeter变量里
vars.put("putdata", JSONObject.valueToString(putdata));
运行的结果（不是用上面的参数值运行，是我脚本实例的结果，但数据结构是一样的），
putdata的值：putdata={"671":{"id":"2660","percent":"25","value":"99"},"672":{"id":"2661","percent":"25","value":"99"},"673":{"id":"2662","percent":"25"},"670":{"id":"2655","percent":"25"}}，这个值被作为变量“putdata”给jmeter使用，我们可以在debugSampler看到这个值：
一些tips
1. 新建处理器的时候，注意选择前置还是后置，前置是指在请求之前进行处理；后置是指在请求之后进行处理。
2. 调试脚本，可以新建一个debugSampler，脚本运行之后，我们可以在查看结果树里查看他的响应数据来观察变量的值；如果是beanshell，可以先在编译器里调试好再复制进去，或者直接打包成jar引用。
