产品实验局环境同事发现分布图与实际业务对不上，因为页面展示流量较大的RTSP，客户并没有相关APP业务。起初以为是前端业务逻辑出问题了：
查看前端数据库查询代码，逻辑十分简单，从数据库中获取app字段的值，然后对应app字典，前端展示：
APP_DICT = get_app_dict()result = {"list": {}}if len(data) != 0:for i in range(len(data)): ????if data[i][1] in APP_DICT: ??????????result["list"][APP_DICT[data[i][1]]] = data[i][0]部分字典示例：
???<platform display_name="RTSP" id="1" name="RTSP" name_en_US="RTSP" name_zh_CN="RTSP" name_zh_TW="RTSP"> ?????<app category="media" display_name="RTSP" id="9209509394251777" name="RTSP" name_en_US="RTSP" name_zh_CN="RTSP" name_zh_TW="RTSP" risk="1" subcategory="photo-video" tags="exband|tunnel|widely" technology="cs"/> ???</platform> ???<platform display_name="联众" id="2" name="联众" name_en_US="Ourgame" name_zh_CN="联众" name_zh_TW="聯眾"> ?????<app category="media" display_name="联众游戏" id="1354598325420033" name="联众游戏" name_en_US="Ourgame" name_zh_CN="联众游戏" name_zh_TW="聯眾遊戲" risk="1" subcategory="game" tags="exband|widely" technology="bs"/> ?????<app category="co" display_name="联众好友" id="1354598325420034" name="联众好友" name_en_US="OurFriend" name_zh_CN="联众好友" name_zh_TW="聯眾好友" risk="1" subcategory="im" tags="
前端没问题，肯定是数据逻辑统计有问题，查看hive中数据正常，kafka中的原始数据也正常，猜想有可能是写postgresql中的逻辑出现问题了。
数据库这个表其他字段都没有错，就是这个字段取值错了：
id | alert_time | process_time | protocol | app | app_proto | bytesall | packetsall
---------+------------+--------------+----------+-----+-----------+------------+------------
?897737 | 1527704580 | ??1527704640 | ???????6 | ??2 | ??????100 | ??????7150 | ????????32
?897772 | 1527704520 | ??1527704700 | ???????6 | ??0 | ????????8 | ??????6966 | ???????112
?897773 | 1527704640 | ??1527704700 | ???????6 | ??0 | ????????8 | ???1344022 | ??????2465
?897774 | 1527704580 | ??1527704700 | ???????6 | ??0 | ????????8 | ????232312 | ???????864
?897775 | 1527704340 | ??1527704700 | ???????6 | ??0 | ????????8 | ?????56384 | ???????166
?897776 | 1527704640 | ??1527704700 | ??????17 | ??4 | ???????31 | ???????552 | ?????????6
?897777 | 1527704700 | ??1527704700 | ???????6 | ??0 | ????????8 | ????107814 | ???????776
?897778 | 1527704400 | ??1527704700 | ???????6 | ??0 | ????????8 | ?????12524 | ????????52
查看数据写数据库的代码，发现app这个值定义int，之前如果数据超过int，应该是溢出为负值了，为什么给了一个0-10的整数值呢？
JSONObject newObject = new JSONObject();
newObject.put("tablename", key);
for (String[] field : fieldList) {
???fieldName = field[0];
???fieldType = field[1];
???if(fieldType.equals("String")){
???????newObject.put(fieldName,obj.getString(fieldName));
???}else if(fieldType.equals("long")){
???????newObject.put(fieldName,obj.getLong(fieldName));
???}else{
???????newObject.put(fieldName,obj.getInteger(fieldName));
???}
}
return newObject.toJSONString().getBytes();
这里通过jsonobject 获取参数的值，如果是int类型的，直接根据getInteger获取值：
public class Test { ???public static void main(String[] args) { ???????String data = "{app:16026481486462977}"; ???????JSONObject newObject = JSON.parseObject(data); ???????System.out.println(newObject); ???????System.out.println(newObject.getInteger("app")); ???}}定义了一个app的正常值，通过getInteger函数获取，果然返回了1：
{"app":16026481486462977}1 继续查看这个函数的方法：
???public Integer getInteger(String key) { ???????Object value = get(key); ???????return castToInt(value); ???}发现调用了castToInt方法，应该是直接返回了int值，继续看castToInt：
???public static Integer castToInt(Object value) { ???????if (value == null) { ???????????return null; ???????} ???????if (value instanceof Integer) { ???????????return (Integer) value; ???????} ???????if (value instanceof Number) { ???????????return ((Number) value).intValue(); ???????} ???????if (value instanceof String) { ???????????String strVal = (String) value; ???????????if (strVal.length() == 0 // ???????????????|| "null".equals(strVal) // ???????????????|| "NULL".equals(strVal)) { ???????????????return null; ???????????} ???????????????????????if (strVal.indexOf(‘,‘) != 0) { ???????????????strVal = strVal.replaceAll(",", ""); ???????????} ???????????return Integer.parseInt(strVal); ???????} ???????if (value instanceof Boolean) { ???????????return ((Boolean) value).booleanValue() ? 1 : 0; ???????} ???????throw new JSONException("can not cast to int, value : " + value); ???}这个方法实际是解决了所有异常场景，如果为int直接返回，如果为其他Number类型的值，直接调用Number 类的intValue方法：
查看JDK对该方法的描述，是强制将Number类型的值转为了int：
到这里原因应该是找到了，改为getLong，修改数据库该字段的类型为Bigint类型，再次启动入库的spark JOB：
id | alert_time | process_time | protocol | app | app_proto | bytesall | packetsall
----+------------+--------------+----------+-------------------+-----------+-----------+------------
?1 | 1528426860 | ??1528426980 | ??????17 | ?2040693581152257 | ????????0 | ??????990 | ????????11
?2 | 1528426920 | ??1528426980 | ??????17 | ?2040693581152257 | ????????0 | ?????1170 | ????????13
?3 | 1528426860 | ??1528426980 | ??????17 | ?4582764464570369 | ???????21 | ??????411 | ?????????5
?4 | 1528426920 | ??1528426980 | ??????17 | ?4582764464570369 | ???????21 | ??????160 | ?????????2
?5 | 1528426860 | ??1528426980 | ??????17 | 16026481486462977 | ????????0 | ??????184 | ?????????2
?6 | 1528426860 | ??1528426980 | ???????6 | ?4037406697193473 | ???????20 | ???????92 | ????????27
?7 | 1528426560 | ??1528426980 | ???????6 | ?3087428650795009 | ????????8 | ?????9128 | ????????33
?8 | 1528426920 | ??1528426980 | ???????6 | ?4037406697193473 | ???????20 | ??????414 | ????????83
?9 | 1528426920 | ??1528426980 | ???????6 | ?3087428650795009 | ????????8 | ????24654 | ????????78
10 | 1528426860 | ??1528426980 | ???????6 | ?3087428650795009 | ????????8 | ????68500 | ???????122
11 | 1528426920 | ??1528426980 | ??????17 | ?6157265115545601 | ????????0 | ?????7888 | ???????102
12 | 1528426920 | ??1528426980 | ???????6 | 27637324275777538 | ????????0 | ??????307 | ????????10
13 | 1528423020 | ??1528426980 | ???????6 | ?4037406697193473 | ???????20 | ???????46 | ?????????8
取值正常，bug得以发现并解决。
Jsonobject 类getInteger函数取值异常
原文地址：https://www.cnblogs.com/lihao7/p/9155018.html