java-mongodb查询笔记

一.模糊查询

1.只能是字符串

2.代码如下:

String tableName = "bidding_notice";

String dbName = "spider";

BasicDBObject qryParam = new BasicDBObject();

qryParam.put(SpiderMapKeyConst.KEY_TASK_ID, "82");

BasicDBObject regex = new BasicDBObject();

regex.put("$regex", ".*?暂无数据.*?");

qryParam.put(SpiderMapKeyConst.KEY_CONTENT, regex);

long count = mongoService.count(dbName, tableName, qryParam);

关键在于正则

regex.put("$regex", ".*?暂无数据.*?"

二.聚合查询

格式

{ $group : { _id : < expression > , < field1 >: { < accumulator1 > : < expression1 > }, ... } }

例子

db.getCollection('bidding_notice').aggregate([ { "$match" :{ "siteId" :"346"}},{ "$group" : { "_id" : "$siteId" , "count" : { "$sum" : 1}}}])

三.$ne详解

---

This will return all documents with a key called "IMAGE URL", but they may still have a null value.

db . mycollection . find ({ "IMAGE URL" :{ $exists : true }});

This will return all documents with both a key called "IMAGE URL"   and   a non-null value.

db . mycollection . find ({ "IMAGE URL" :{ $ne : null }});

Also, according to the docs, $exists currently can't use an index, but $ne can.

Edit: Adding some examples due to interest in this answer

Given these inserts:

db . test . insert ({ "num" : 1 , "check" : "check value" });

db . test . insert ({ "num" : 2 , "check" : null });

db . test . insert ({ "num" : 3 });

This will return all three documents:

db . test . find ();

This will return the first and second documents only:

db . test . find ({ "check" :{ $exists : true }});

This will return the first document only:

db . test . find ({ "check" :{ $ne : null }});

This will return the second and third documents only:

db . test . find ({ "check" : null })

---

三.根据字符串长度筛选

mongodb命令:

db.getCollection('bidding_notice').count({"taskId":"82","content": {$ne: null}, $where:"this.content.length< 400"})

java代码:

BasicDBObject qryParam = new BasicDBObject();

BasicDBObject content = new BasicDBObject();

content.put("$ne", null);

qryParam.put(SpiderMapKeyConst.KEY_CONTENT, content);

qryParam.put("$where", "this.content.length < 400");

---

四.根据id查询

自定义id的时候,并不符合new OjbectId()的规定,所以要根据id查询的时候需要迂回一下

java代码:

BasicDBObject query = new BasicDBObject();

query.put("_id", id);

---

五.$not在Java代码的实现

$ne不起作用的时候,就需要用$not,但是$not不能直接接$regex

mongodb查询代码:

db.getCollection('bidding_notice').count({ "siteId" : "97", "parsed":4, "title" : { $not : /^.*?废标|失败|终止|中止|作废公告|示$/}})

我用的是3.4.2版本的mongo-driver+dubbo,消费者调用的时候有个坑,dubbo远程调用,参数一定要实现serialize接口,而mongoclient的filters是没有实现的,bson又是一个接口,还有java的pattern也无法通过rpc的decode

正常调用(本地):

Pattern pattern = Pattern.compile ("^.*?废标|失败|终止|中止|作废公告|示$");

Bson bson = new BasicDBObject("title", pattern);

Bson title = Filters.not(bson);

// Bson title = new BasicDBObject(QueryOperators.NOT, bson);

 Bson and = Filters.and(title, new BasicDBObject("siteId", "97"), new BasicDBObject("parsed", 4));

mongoService.count(mongoService.getDefaultDbName(), tableName, and);

远程调用(RPC):

Map<String, Object> query = Maps.newHashMap();

Map<String, Object> title = Maps.newHashMap(); // 注意这里与上面的区别,上面的可以直接用pattern,这里decode有问题

title.put("title", "^.*?废标|失败|终止|中止|作废|流标公告|示$");

query.put("$not", title);

List<Bson> queryList = Lists.newArrayList();

Map patternMap = (Map) value;

Set set = patternMap.keySet();

for (Object key : set) {

    String regexStr = MapUtils.getString(patternMap, key);

    Pattern pattern = Pattern.compile(regexStr );

    Bson bson = new BasicDBObject((String) key, pattern);

    queryList.add(Filters.not(bson));

}

mongoService.count(mongoService.getDefaultDbName(), tableName, Filters.and(queryList); );

---