java Elastic 客戶端基本使用
引入jar
compile 'org.elasticsearch:elasticsearch:5.5.0'
compile 'org.elasticsearch.client:transport:5.5.0
client基本使用
得到client
Settings settings = Settings.builder().put("cluster.name", "lw-6-test").build();
TransportClient client = new PreBuiltTransportClient(settings);
client.addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.10.10.6"), 9300));
關(guān)閉資源
client.close();
- 搜索關(guān)鍵字全部要小寫。
get得到指定index type id的數(shù)據(jù)
public static void prepareGet(TransportClient client) throws Exception {
GetResponse response = client.prepareGet("mytest", "test", "p1").get();
System.out.println(response);
}
output:
{"_index":"mytest","_type":"test","_id":"p1","_version":1,"found":true,"_source":{"name":"mac Book 筆記本",
"price":1233,
"description":"這是筆記本",
"cats":["3c","computer"]
}
}
insert添加數(shù)據(jù)
public static void insert(TransportClient client) throws Exception {
Map<String,Object> resource = new HashMap<>();
resource.put("name","mac Note");
resource.put("price",8877);
resource.put("description","mac Note 新款");
IndexRequestBuilder index = client.prepareIndex("mytest", "test");
IndexResponse insertResponse = index.setSource(resource).execute().get();
System.out.println(insertResponse);
}
output:
IndexResponse[index=mytest,type=test,id=AV8CZmTgGnilLCUrybiV,version=1,result=created,shards={"total":2,"successful":1,"failed":0}]
delete刪除數(shù)據(jù)
public static void delete(TransportClient client) throws Exception{
BulkByScrollResponse response = DeleteByQueryAction.INSTANCE.newRequestBuilder(client)
.filter(QueryBuilders.matchQuery("name", "mac")) //搜索
.source("mytest") //index
.get();
long deleted = response.getDeleted();
System.out.println("刪除個(gè)數(shù): "+deleted);
}
update 修改數(shù)據(jù)
public static void update(TransportClient client) throws Exception{
Map<String,Object> data = new HashMap<>();
data.put("name","new mac node");
UpdateRequest updateRequest = new UpdateRequest();
updateRequest.index("mytest");
updateRequest.type("test");
updateRequest.id("AV8CfcSLGnilLCUryoEl");
updateRequest.doc(data);
UpdateResponse response = client.update(updateRequest).get();
System.out.println(response);
}
output:
UpdateResponse[index=mytest,type=test,id=AV8CfcSLGnilLCUryoEl,version=2,result=updated,shards=ShardInfo{total=2, successful=1, failures=[]}]
MultiGet查詢多個(gè)index
public static void multiIndex(TransportClient client) throws Exception {
MultiGetResponse multiGetItemResponses = client.prepareMultiGet()
.add("mytest","test","AV8CfcSLGnilLCUryoEl") //多個(gè)index
.add("instestdb_business_log-2017.09","instestdb_business_log","AV7KHPtGDF9uyeK_lXln") //多個(gè)index
.get();
for(MultiGetItemResponse itemResponses : multiGetItemResponses) {
GetResponse response = itemResponses.getResponse();
if(response.isExists()) {
String json = response.getSourceAsString(); //獲取到_source field
System.out.println(json);
}
}
}
Bulk API 一次請(qǐng)求多個(gè)添加和刪除
public static void BulkRequest(TransportClient client) throws Exception {
BulkRequestBuilder bulkRequest = client.prepareBulk();
IndexRequestBuilder index1 = client.prepareIndex("mytest", "test");
IndexRequestBuilder index2 = client.prepareIndex("mytest", "test");
Map<String,Object> resource = new HashMap<>();
resource.put("name","華碩");
resource.put("price",5577);
resource.put("description","華碩z460");
index1.setSource(resource);
Map<String,Object> resource1 = new HashMap<>();
resource1.put("name","小米2");
resource1.put("price",4577);
resource1.put("description","新機(jī)超薄");
index1.setSource(resource);
index2.setSource(resource1);
bulkRequest.add(index1);
bulkRequest.add(index2);
BulkResponse bulkResponse = bulkRequest.get();
if(bulkResponse.hasFailures()) {
System.out.println(bulkResponse.buildFailureMessage());
}
bulkResponse.forEach(response ->{
System.out.println(response.getId());
});
}
query dsl 使用
import static org.elasticsearch.index.query.QueryBuilders.*;
Settings settings = Settings.builder()
.put("cluster.name", "lw-6-test").build();
TransportClient client = new PreBuiltTransportClient(settings);
client.addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.10.10.6"), 9300));
allquery(client); //具體的query dsl查詢
client.close();
Match All Query 查詢所有的數(shù)據(jù)
public static void allquery(TransportClient client) throws Exception{
QueryBuilder qb = matchAllQuery();
SearchResponse response = client.prepareSearch("mytest").setTypes("test").setSize(3).setQuery(qb).get();
System.out.println("length: "+response.getHits().getHits().length );
if(response.getHits().getTotalHits() != 0) {
for (SearchHit hit : response.getHits().getHits()) {
System.out.println(hit.getSourceAsString());
}
}
}
Match Query 查詢單一條件的數(shù)據(jù)
public static void myMatchQuery(TransportClient client) throws Exception {
QueryBuilder qb = matchQuery("name","mac");
SearchResponse response = client.prepareSearch("mytest").setTypes("test").setQuery(qb).get();
System.out.println("length: "+response.getHits().getHits().length );
if(response.getHits().getTotalHits() != 0) {
for (SearchHit hit : response.getHits().getHits()) {
System.out.println(hit.getScore()+" --> "+hit.getSourceAsString());
}
}
}
MultiMatchQuery 在多個(gè)字段中查詢一個(gè)關(guān)鍵字
QueryBuilder qb = multiMatchQuery("mac","description","name"); //mac是要搜索的詞 description，name 都是字段
Common Terms Query 搜索term
public static void myCommonTermsQuery(TransportClient client ) throws Exception{
QueryBuilder qb = commonTermsQuery("description","mac");
print(qb,client);
}
Simple Query String Query 簡(jiǎn)單字符串查詢可以使用正則
public static void mySimpleQueryString(TransportClient client ) {
QueryBuilder qb = queryStringQuery("mac*^2").field("name");
print(qb,client);
}
term 搜索關(guān)鍵詞一個(gè)
public static void myTermQuery(TransportClient client) {
QueryBuilder qb = termQuery("name","mac2");
print(qb,client);
}
terms 搜索關(guān)鍵詞多個(gè)
QueryBuilder qb = termsQuery("name_str","小米","戴爾");
print2(qb,client);
range query 范圍搜索
public static void myRangeQuery(TransportClient client){
QueryBuilder qb = rangeQuery("price").from(3399)
.to(6399)
.includeLower(true)
.includeUpper(false);
print2(qb,client);
}
QueryBuilder qb = rangeQuery("price").gte(3399).lt(6399);
Exists Query 查找字段是否存在 存在則返回所有的數(shù)據(jù)，不存在返回0
public static void myExistsQuery(TransportClient client) {
QueryBuilder qb = existsQuery("name_str");
print2(qb,client);
}
Wildcard Query 通配符查詢
QueryBuilder qb = wildcardQuery("user", "k?mc*");
?
Regexp Query支持正則表達(dá)式的查詢
QueryBuilder qb = regexpQuery("name.first", "s.*y");
Fuzzy Query 模糊查詢
QueryBuilder qb = fuzzyQuery(
"name",
"kimzhy"
);
ids Query 根據(jù)id 查詢
QueryBuilder qb = idsQuery("my_type", "type2")
.addIds("1", "4", "100");
QueryBuilder qb = idsQuery()
.addIds("AV8HhVC8FiG-4m4G2rYp","AV8HhVB6FiG-4m4G2rYm");
?
復(fù)合查詢
Contant Score Query 指定score
QueryBuilder qb = constantScoreQuery(matchQuery("name_str", "聯(lián)想")).boost(3.0f);
Bool Query must mustNot should 查詢
所有的 must 子句必須匹配, 并且所有的 must_not 子句必須不匹配, 但是多少 should 子句應(yīng)
該匹配呢? 默認(rèn)的,不需要匹配任何 should 子句,一種情況例外:如果沒有must子句,就必須至少匹 配一個(gè)should子句。
GET mttest-2017.10.41/web1/_search
{
"query": {
"bool": {
"must": [
{
"match": {
"request": "choosehospital/index"
}
},
{
"match": {
"response": "500"
}
}
]
}
}
}
public static void myBoolQuery(TransportClient client) {
QueryBuilder qb = boolQuery().must(termQuery("name_str","小米"))
.filter(matchQuery("price",3599))
.filter(matchQuery("description","lihao"));
print2(qb, client);
}
QueryBuilder qb = boolQuery()
.must(termQuery("content", "test1"))
.must(termQuery("content", "test4"))
.mustNot(termQuery("content", "test2"))
.should(termQuery("content", "test3"))
.filter(termQuery("content", "test5"));
?
indices query查詢多個(gè)index
用來(lái)查詢多個(gè)index,對(duì)于指定內(nèi)的index,執(zhí)行指定的內(nèi)部query;對(duì)于指定外的index,執(zhí)行
no_match_query設(shè)定的條件
private static void print(QueryBuilder qb, TransportClient client) {
SearchResponse response = client.prepareSearch("mytest").setTypes("test").setQuery(qb).get();
System.out.println("length: " + response.getHits().getHits().length);
if (response.getHits().getTotalHits() != 0) {
for (SearchHit hit : response.getHits().getHits()) {
System.out.println(hit.getScore() + " --> " + hit.getSourceAsString());
}
}
}
private static void print2(QueryBuilder qb, TransportClient client) {
SearchResponse response = client.prepareSearch("mytest_1").setTypes("test").setQuery(qb).get();
System.out.println("length: " + response.getHits().getHits().length);
if (response.getHits().getTotalHits() != 0) {
for (SearchHit hit : response.getHits().getHits()) {
System.out.println(hit.getScore() + " --> " + hit.getSourceAsString());
}
}
}
scroll分頁(yè)
public static void main(String ...arg) throws Exception {
//鏈接服務(wù)器
Settings settings = Settings.builder()
.put("cluster.name","lw-6-test").build();
TransportClient client = new PreBuiltTransportClient(settings);
client.addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.10.10.6"), 9300));
//設(shè)置搜索條件
QueryBuilder qb = termQuery("name_str","筆記本");
// 按照price 降序 每次查詢2條 第一次不需要設(shè)置sroll scrollid
SearchResponse scrollResp = client.prepareSearch("mytest_1").setTypes("test")
.addSort("price", SortOrder.DESC)
.setScroll(new TimeValue(30000))
.setQuery(qb)
.setSize(2).get();
System.out.println("length: " + scrollResp.getHits().getHits().length);
int count = 1;
do{
System.out.println("第 " +count+ " 次");
System.out.println();
for (SearchHit hit : scrollResp.getHits().getHits()){
System.out.println(hit.getScore() + " --> " +hit.getSourceAsString());
}
System.out.println("scrollid: "+scrollResp.getScrollId());
//設(shè)置sroll id
scrollResp =client.prepareSearchScroll(scrollResp.getScrollId()).setScroll(new TimeValue(60000)).execute().actionGet();
System.out.println();
count++;
} while (scrollResp.getHits().getHits().length !=0);
client.close();
}
?
prepareMultiSearch多個(gè)條件查詢
public class MultiSearchDemo {
public static void main(String ...arg) throws Exception{
Settings settings = Settings.builder().put("cluster.name", "lw-6-test").build();
TransportClient client = new PreBuiltTransportClient(settings);
client.addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.10.10.6"), 9300));
QueryBuilder query1 = termQuery("name_str","小米");
QueryBuilder query2 = termQuery("name_str","戴爾");
SearchRequestBuilder srb1 = client.prepareSearch("mytest_1").setTypes("test").setQuery(query1);
SearchRequestBuilder srb2 = client.prepareSearch("mytest_1").setTypes("test").setQuery(query2);
MultiSearchResponse sr = client.prepareMultiSearch().add(srb1).add(srb2).get();
long nbHits =0;
for(MultiSearchResponse.Item item : sr.getResponses()) {
SearchResponse response = item.getResponse();
nbHits += response.getHits().getTotalHits();
if(response.getHits().getHits().length >0) {
for(SearchHit hit : response.getHits().getHits()) {
System.out.println(hit.getScore()+" -----> "+hit.getSourceAsString());
}
}
System.out.println("-------------------------");
}
System.out.println(nbHits);
client.close();
}
}
聚合
概述
ES 的聚合框架提供對(duì)查詢得到的數(shù)據(jù)進(jìn)行分組和匯總統(tǒng)計(jì)，以提供復(fù)雜的統(tǒng)計(jì)分析功能。
ES支持在一次聚合查詢中，可以同時(shí)得到聚合的具體結(jié)果并再次進(jìn)行聚合，也就是聚合是可以嵌套的。
這非常有用,你可以通過(guò)一次操作得到多次聚合的結(jié)果,從而避免多次請(qǐng)求,減少網(wǎng)絡(luò)和服務(wù)器的負(fù)擔(dān)。
聚合的類型
1：Bucketing（桶）聚合：劃分不同的“桶”，將數(shù)據(jù)分配到不同的“桶” 里，然后再進(jìn)行聚合，非常類似sql 中的group 語(yǔ)句的含義。
2：Metric(指標(biāo))聚合：指標(biāo)聚合主要針對(duì)number類型的數(shù)據(jù)，在一組文檔中，保持對(duì)要聚合的指標(biāo)的跟蹤和計(jì)算，需要ES做比較多的計(jì)算工作。
3：Pipeline(管道)聚合：用來(lái)聚合其它聚合輸出的結(jié)果以及相關(guān)指標(biāo)
聚合的基本語(yǔ)法
"aggregations" : { //定義聚合對(duì)象,也可用 "aggs"
"<aggregation_name>" : { //聚合的名稱,用戶自定義
"<aggregation_type>" : { //聚合類型,比如 "histogram" "avg"
<aggregation_body>
}
[,"meta" : { [<meta_data_body>] } ]?
[,"aggregations" : { [<sub_aggregation>]+ } ]?
}
[,"<aggregation_name_2>" : { ... } ]* ////定義額外的多個(gè)平級(jí)聚合,只有Bucketing類型才有意義
}
?
GET mytest_1/test/_search
{
"aggs" : {
"avg_price" : { "avg" : { "field" : "price" } }
}
}
output
{ "aggregations": {
"avg_price": {
"value": 4954.555555555556
}
}
?
?
Metric 使用
求平均值
public static void main(String ...arg) throws Exception {
Settings settings = Settings.builder().put("cluster.name", "lw-6-test").build();
TransportClient client = new PreBuiltTransportClient(settings);
client.addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.10.10.6"),9300));
avg(client);
client.close();
}
private static void avg(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
SearchResponse sr = search.addAggregation(AggregationBuilders.avg("avg_price").field("price")).execute().actionGet();
Avg result = sr.getAggregations().get("avg_price");
System.out.println(result.getValue());
}
POST mytest_1/test/_search?size=0
{
"aggs" : {
"avg_price" : {
"avg" : { "field" : "price" }
}
}
}
POST mytest_1/test/_search?size=0
{
"aggs" : {
"all_cats" : {
"terms" : { "field" : "tag.keyword" },
"aggs" : {
"avg_price" : {
"avg" : { "field" : "price" }
}
}
}
}
}
"aggregations": {
"all_cats": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "筆記本",
"doc_count": 8,
"avg_price": {
"value": 5124
}
},
{
"key": "聯(lián)想",
"doc_count": 4,
"avg_price": {
"value": 5649
}
},
{
"key": "小米",
"doc_count": 2,
"avg_price": {
"value": 4399
}
},
{
"key": "惠普",
"doc_count": 1,
"avg_price": {
"value": 2399
}
},
{
"key": "戴爾",
"doc_count": 1,
"avg_price": {
"value": 7199
}
}
]
}
}
分類求取平均值
PUT mytest_1/_mapping/test
{
"properties": {
"tag": {
"type": "text",
"fielddata": true
}
}
}
private static void avg1(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setSize(0).setTypes("test");
TermsAggregationBuilder tag = AggregationBuilders.terms("tags").field("tag.keyword");
AvgAggregationBuilder price = AggregationBuilders.avg("avg_price").field("price");
tag.subAggregation(price);
SearchResponse sr = search.addAggregation(tag).execute().actionGet();
System.out.println(sr);
}
Cardinality 去除重復(fù)數(shù)據(jù)
用來(lái)對(duì)單個(gè)數(shù)據(jù)進(jìn)行匯總,計(jì)算不重復(fù)的值的數(shù)量。
public static void main(String... arg) throws Exception {
Settings settings = Settings.builder().put("cluster.name", "lw-6-test").build();
TransportClient client = new PreBuiltTransportClient(settings);
client.addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("10.10.10.6"), 9300));
cardinality(client);
client.close();
}
private static void cardinality(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
SearchResponse sr = search.addAggregation(AggregationBuilders.cardinality("type_count").field("price")).execute().actionGet();
Cardinality result = sr.getAggregations().get("type_count");
System.out.println("type_count: "+result.getValue());
}
POST mytest_1/test/_search?size=0
{
"aggs" : {
"type_count" : {
"cardinality" : {
"field" : "price"
}
}
}
}
out:
{
"took": 7,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 8,
"max_score": 0,
"hits": []
},
"aggregations": {
"type_count": {
"value": 6
}
}
}
?
?
Stats 聚合操作 count min max avg sum
private static void stats(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
SearchResponse sr = search.addAggregation(AggregationBuilders.stats("price_stats").field("price")).execute().actionGet();
Stats stats = sr.getAggregations().get("price_stats");
System.out.println(stats.getAvgAsString());
System.out.println(stats.getMaxAsString());
System.out.println(stats.getMinAsString());
System.out.println(stats.getSumAsString());
}
POST mytest_1/test/_search?size=0
{
"aggs" : {
"price_stats" : { "extended_stats" : { "field" : "price" } }
}
}
out:
"aggregations": {
"price_stats": {
"count": 8,
"min": 2399,
"max": 7199,
"avg": 5124,
"sum": 40992,
"sum_of_squares": 231958008,
"variance": 2739375,
"std_deviation": 1655.1057368035433,
"std_deviation_bounds": {
"upper": 8434.211473607087,
"lower": 1813.7885263929134
}
}
}
?
Percentiles 百分比 查看一個(gè)百分比對(duì)應(yīng)的值
這是一個(gè)多值的指標(biāo)聚集,用來(lái)計(jì)算聚合文檔中,在某個(gè)百分比或某個(gè)區(qū)間,所對(duì)應(yīng)的觀測(cè)值,
1:缺省的percentile的區(qū)間是[ 1, 5, 25, 50, 75, 95, 99 ]。
2:觀測(cè)值通常都是近似的,有很多不同的算法來(lái)計(jì)算。
?
? 例如:第九十五個(gè)百分值是大于所觀察到的值的95%的值。
POST mytest_1/test/_search?size=0
{
"aggs" : {
"price_percent" : {
"percentiles" : {
"field" : "price"
}
}
}
}
POST mytest_1/test/_search?size=0
{
"aggs" : {
"price_percent" : {
"percentiles" : {
"field" : "price" ,
"percents" : [0.1,50,95, 99, 100] //自定義百分比區(qū)間
}
}
}
}
"aggregations": {
"price_percent": {
"values": {
"1.0": 2468.9999999999995,
"5.0": 2749,
"25.0": 3549,
"50.0": 5799,
"75.0": 6399,
"95.0": 6918.999999999999,
"99.0": 7143
}
private static void percentile(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
SearchResponse sr = search.addAggregation(AggregationBuilders.percentiles("price_percent").field("price")).execute().actionGet();
Percentiles percentile = sr.getAggregations().get("price_percent");
System.out.println(percentile.percentileAsString(80));
}
//自定義百分比區(qū)間
private static void percentile2(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
SearchResponse sr = search.addAggregation(AggregationBuilders.percentiles("price_percent").percentiles(0.1,50,95, 99, 100).field("price")).execute().actionGet();
System.out.println(sr);
Percentiles percentile = sr.getAggregations().get("price_percent");
System.out.println(percentile.percentileAsString(80));
}
?
Value Count 計(jì)算聚合值的數(shù)量
POST mytest_1/test/_search?size=0
{
"aggs" : {
"types_count" : { "value_count" : { "field" : "price" } }
}
}
out:
"aggregations": {
"types_count": {
"value": 8
}
private static void valueCount(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
SearchResponse sr =search.addAggregation(AggregationBuilders.count("value_count").field("price")).execute().actionGet();
ValueCount valueCount = sr.getAggregations().get("value_count");
System.out.println(valueCount.getValue());
}
?
TOP hits
用來(lái)取符合條件的前n條數(shù)據(jù)。 包含的選項(xiàng)有:from、size、sort。
private static void topHits(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test");
TopHitsAggregationBuilder addtion = AggregationBuilders.topHits("top_price_hits").sort("price", SortOrder.DESC).fieldDataField("price")
.size(5);
SearchResponse sr =search.addAggregation(addtion).execute().actionGet();
TopHits topHits = sr.getAggregations().get("top_price_hits");
System.out.println();
SearchHit[] hits = topHits.getHits().internalHits();
for(SearchHit searchHit : hits) {
System.out.println(searchHit.getSourceAsString());
}
}
?
?
bucket 使用
Histogram
條形圖聚合，根據(jù)文檔中的謀改革字段來(lái)分組。一個(gè)文檔屬于某個(gè)通，計(jì)算過(guò)程大致如下：
rem = value % interval
if (rem < 0) {
rem += interval
}
bucket_key = value - rem
1:可配置的參數(shù):
(1)field:字段,必須為數(shù)值類型
(2)interval:分桶間距 (3)min_doc_count:最少文檔數(shù),桶過(guò)濾,只有不少于這么多文檔的桶才會(huì)返回
(4)extended_bounds:范圍擴(kuò)展
(5)order:對(duì)桶排序,如果 histogram 聚合有一個(gè)權(quán)值聚合類型的“直接”子聚合,那么排序可以使用 子聚合中的結(jié)果
(6)offset:桶邊界位移,默認(rèn)從0開始 (7)keyed:hash結(jié)構(gòu)返回,默認(rèn)以數(shù)組形式返回每一個(gè)桶 (8)missing:配置缺省默認(rèn)值
POST mytest_1/test/_search?size=0
{
"aggs" : {
"prices" : {
"histogram" : {
"field" : "price",
"interval" : 2000
}
}
}
}
out:
"aggregations": {
"prices": {
"buckets": [
{
"key": 2000,
"doc_count": 3
},
{
"key": 4000,
"doc_count": 1
},
{
"key": 6000,
"doc_count": 4
}
]
}
private static void histogram(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test").setSize(0);
HistogramAggregationBuilder addtion = AggregationBuilders.histogram("prices").interval(2000).field("price");
SearchResponse sr = search.addAggregation(addtion).execute().actionGet();
Histogram histogram = sr.getAggregations().get("prices");
histogram.getBuckets().forEach(bucket->{
System.out.println(bucket.getKeyAsString()+" ----> "+bucket.getDocCount());
});
}
Range
范圍聚合，是對(duì)某個(gè)字段的值，按照設(shè)定的范圍進(jìn)行分組。
POST mytest_1/test/_search?size=0
{
"aggs" : {
"price_ranges" : {
"range" : {
"field" : "price",
"ranges" : [
{ "to" : 3000 },
{ "from" : 3000, "to" : 5000 },
{ "from" : 5000 }
]
}
}
}
}
out:
"aggregations": {
"price_ranges": {
"buckets": [
{
"key": "*-3000.0",
"to": 3000,
"doc_count": 1
},
{
"key": "3000.0-5000.0",
"from": 3000,
"to": 5000,
"doc_count": 2
},
{
"key": "5000.0-*",
"from": 5000,
"doc_count": 5
}
]
}
private static void range(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test").setSize(0);
AggregationBuilder addtion = AggregationBuilders.range("price_ranges").field("price")
.addUnboundedTo(3000)
.addRange(3000,5000)
.addUnboundedFrom(5000);
SearchResponse sr = search.addAggregation(addtion).execute().actionGet();
Range histogram = sr.getAggregations().get("price_ranges");
histogram.getBuckets().forEach(bucket->{
String key = bucket.getKeyAsString();
String from = bucket.getFromAsString();
String to = bucket.getToAsString();
long count = bucket.getDocCount();
System.out.println("key : "+key+"\t form: "+from+"\t to:"+to+"\t count:"+count);
});
}
Date Range
Ip Range
Terms
詞元聚合，以指定的字段內(nèi)的每一個(gè)不重復(fù)的term來(lái)分組，并計(jì)算每個(gè)組內(nèi)文檔的個(gè)數(shù)。
POST mytest_1/test/_search?size=0
{
"aggs" : {
"all_cats" : {
"terms" : { "field" : "tag.keyword" }
}
}
}
Filters 過(guò)濾
多過(guò)濾聚合，用多個(gè)過(guò)濾條件，來(lái)對(duì)當(dāng)前文檔進(jìn)行過(guò)濾的聚合，每個(gè)過(guò)濾都包含所有滿足它的文檔，
多個(gè)bucket中可能重復(fù)。
POST mytest_1/test/_search?size=0
{
"size": 0,
"aggs" : {
"messages" : {
"filters" : {
"filters" : {
"filter1" : { "match" : { "tag" : "小米" }},
"filter2" : { "match" : { "tag" : "戴爾" }}
}
}
}
}
}
out:
"aggregations": {
"messages": {
"buckets": {
"filter1": {
"doc_count": 2
},
"filter2": {
"doc_count": 1
}
}
}
private static void filters(TransportClient client) {
SearchRequestBuilder search = client.prepareSearch("mytest_1").setTypes("test").setSize(0);
AggregationBuilder aggregation = AggregationBuilders.filters("filters",
new FiltersAggregator.KeyedFilter("xiaomi",matchQuery("tag","小米")),
new FiltersAggregator.KeyedFilter("daier",matchQuery("tag","戴爾"))
);
SearchResponse sr = search.addAggregation(aggregation).execute().actionGet();
System.out.println(sr);
Filters agg =sr.getAggregations().get("filters");
agg.getBuckets().forEach(entry->{
String key = entry.getKeyAsString(); // bucket key
long docCount = entry.getDocCount();
System.out.println("key : "+key+"\t docCount: "+docCount);
});
}
技術(shù)點(diǎn)
分詞
分詞安裝
https://github.com/medcl/elasticsearch-analysis-ik/tree/v5.5.0
測(cè)試分詞
GET _analyze
{
"analyzer": "ik_smart",
"text": ["中華人民共和國(guó)"]
}
參數(shù)
boost:權(quán)重,缺省1.0
bool
bool 過(guò)濾 :可以用來(lái)合并多個(gè)過(guò)濾條件查詢結(jié)果的布爾邏輯,它包含一下操作符:
must:多個(gè)查詢條件的完全匹配,相當(dāng)于 and。
must_not: 多個(gè)查詢條件的相反匹配,相當(dāng)于 not。
should: 至少有一個(gè)查詢條件匹配, 相當(dāng)于 or。
?
參考文檔
https://www.elastic.co/guide/en/elasticsearch/client/java-api/current/_structuring_aggregations.html
