聚合查詢不是直接查詢文檔數(shù)據(jù)系羞，而是對文檔數(shù)據(jù)按照某些維度進(jìn)行統(tǒng)計(jì)柑肴，如果你熟悉 MySql 的聚合查詢灼芭，這個也就好理解了废膘。之前我們已經(jīng)學(xué)習(xí)了使用 RESTful API 聚合查詢竹海，現(xiàn)在學(xué) Java High Level REST Client 的聚合查詢也就很簡單了，

我們還是使用上一篇的文檔數(shù)據(jù)學(xué)習(xí)聚合查詢：

我們一般可以使用AggregationBuilders類的靜態(tài)方法來構(gòu)建需要的聚合方式丐黄。它會返回一個 Builder 類斋配，當(dāng)然你也可以直接new一個指定聚合方式的 Builder 類。

1、avg

public void avg() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 統(tǒng)計(jì)文檔中age字段的平均值艰争，avgAge相當(dāng)于統(tǒng)計(jì)結(jié)果的名稱
    AvgAggregationBuilder avgBuilder = AggregationBuilders.avg("avgAge").field("age");
    // 設(shè)置聚合查詢
    searchSourceBuilder.aggregation(avgBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Avg avg = response.getAggregations().get("avgAge");
    double value = avg.getValue();
    System.out.println(value);
}

上邊是統(tǒng)計(jì)age的平均值坏瞄，注意，由于沒有添加其它查詢條件甩卓，則會統(tǒng)計(jì)索引中所有文檔鸠匀。

2、max

統(tǒng)計(jì)age的最大值：

public void max() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 統(tǒng)計(jì)文檔中age字段的最大值
    MaxAggregationBuilder maxBuilder = AggregationBuilders.max("maxAge").field("age");
    searchSourceBuilder.aggregation(maxBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Max max = response.getAggregations().get("maxAge");
    double value = max.getValue();
    System.out.println(value);
}

3逾柿、min缀棍、sum

統(tǒng)計(jì)最小值以及求和的實(shí)現(xiàn)上邊的類似，就不詳細(xì)說了：

MinAggregationBuilder minBuilder = AggregationBuilders.min("minAge").field("age");

SumAggregationBuilder sumBuilder = AggregationBuilders.sum("sumAge").field("age");

4机错、range

range表示按區(qū)間統(tǒng)計(jì)爬范，比如指定時間范圍，指定大小區(qū)間等弱匪。如下統(tǒng)計(jì)age在(-∞, 30)青瀑、[30,40]、(40,+∞)三個區(qū)間的人數(shù)：

public void range() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 統(tǒng)計(jì)文檔中age字段的最大值
    RangeAggregationBuilder rangeBuilder = AggregationBuilders.range("rangeAge")
            .field("age")
            .addUnboundedTo(30)
            .addRange(30, 40)
            .addUnboundedFrom(40);
    searchSourceBuilder.aggregation(rangeBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Range range = response.getAggregations().get("rangeAge");
    for (Range.Bucket bucket : range.getBuckets()) {
        // 打印每個區(qū)間的人數(shù)
        System.out.println("age區(qū)間 " + bucket.getKeyAsString() + " 的人數(shù)：" + bucket.getDocCount());
    }

統(tǒng)計(jì)的結(jié)果如下：

5萧诫、filter

filter可以按指定的查詢條件過濾數(shù)據(jù)狱窘，如下統(tǒng)計(jì)姓school是北大的人數(shù)：

public void filter() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 統(tǒng)計(jì)文檔中school是北大的人數(shù)
    // 先構(gòu)建查詢條件
    TermQueryBuilder termQueryBuilder = QueryBuilders.termQuery("school.keyword", "北大");
    // 設(shè)置過濾統(tǒng)計(jì)的查詢條件
    FilterAggregationBuilder filterBuilder = AggregationBuilders.filter("count", termQueryBuilder);
    searchSourceBuilder.aggregation(filterBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Filter filter = response.getAggregations().get("count");
    double value = filter.getDocCount();
    System.out.println(value);
}

6、count

count是統(tǒng)計(jì)數(shù)量的财搁，如下根據(jù)文檔 id 統(tǒng)計(jì)索引中的文檔數(shù)：

public void valueCount() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 根據(jù)文檔id統(tǒng)計(jì)索引的文檔數(shù)
    ValueCountAggregationBuilder valueCountBuilder = AggregationBuilders.count("count").field("_id");
    searchSourceBuilder.aggregation(valueCountBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    ValueCount valueCount = response.getAggregations().get("count");
    double value = valueCount.getValue();
    System.out.println(value);
}

7蘸炸、terms

terms是按指定字段對文檔數(shù)據(jù)進(jìn)行分組，如下按school字段進(jìn)行分組尖奔，統(tǒng)計(jì)出前20組（默認(rèn)10組）搭儒，并按每組的數(shù)據(jù)量升序排列（默認(rèn)降序）：

public void terms() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 按照school分組
    TermsAggregationBuilder termsBuilder = AggregationBuilders.terms("schoolGroup")
            .field("school.keyword")
            // 按每組的數(shù)據(jù)量升序排列
            .order(BucketOrder.aggregation("_count", true))
            // 最多統(tǒng)計(jì)出20組數(shù)據(jù)
            .size(20);
    searchSourceBuilder.aggregation(termsBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Terms terms = response.getAggregations().get("schoolGroup");
    for (Terms.Bucket bucket : terms.getBuckets()) {
        System.out.println(bucket.getKeyAsString() + " 的人數(shù)：" + bucket.getDocCount());
    }
}

8、子統(tǒng)計(jì)

上邊我們使用terms對文檔數(shù)據(jù)按照school字段進(jìn)行了分組提茁，我們還可以對組內(nèi)的數(shù)據(jù)進(jìn)行其它統(tǒng)計(jì)淹禾，例如統(tǒng)計(jì)age的最小值，這就是子統(tǒng)計(jì)茴扁。代碼如下：

public void sub() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 按照school分組
    TermsAggregationBuilder termsBuilder = AggregationBuilders.terms("schoolGroup")
            .field("school.keyword")
            // 按每組的數(shù)據(jù)量升序排列
            .order(BucketOrder.aggregation("_count", true))
            // 最多統(tǒng)計(jì)出20組數(shù)據(jù)
            .size(20)
            // 添加子統(tǒng)計(jì)
            .subAggregation(AggregationBuilders.min("minAge").field("age"));
    searchSourceBuilder.aggregation(termsBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Terms terms = response.getAggregations().get("schoolGroup");
    for (Terms.Bucket bucket : terms.getBuckets()) {
        // 取出子統(tǒng)計(jì)的結(jié)果
        Min min = bucket.getAggregations().get("minAge");
        System.out.println(bucket.getKeyAsString() + " 的人數(shù)：" + bucket.getDocCount() + "铃岔，age的最小值：" + min.getValue());
    }
}

9、topHits

前邊的各種聚合查詢只能統(tǒng)計(jì)出最終的結(jié)果峭火，我們并不能知道那些文檔數(shù)據(jù)參與了統(tǒng)計(jì)毁习，topHits可以用來跟蹤正在參與分組聚合統(tǒng)計(jì)的文檔數(shù)據(jù)，我在前邊terms例子的基礎(chǔ)上繼續(xù)修改卖丸，來跟蹤每組內(nèi)的前20條數(shù)據(jù)（默認(rèn)10條數(shù)據(jù)）纺且，并按age升序排列：

public void topHits() throws IOException {
    SearchRequest request = new SearchRequest("user");
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    // 跟蹤正在參與分組聚合統(tǒng)計(jì)的文檔數(shù)據(jù)
    TopHitsAggregationBuilder topHitsBuilder = AggregationBuilders.topHits("groupData")
            // 跟蹤前20條數(shù)據(jù)
            .size(20)
            // 按age升序排列
            .sort("age", SortOrder.ASC);
    // 按照school分組
    TermsAggregationBuilder termsBuilder = AggregationBuilders.terms("schoolGroup")
            .field("school.keyword")
            // 按每組的數(shù)據(jù)量升序排列
            .order(BucketOrder.aggregation("_count", true))
            // 最多統(tǒng)計(jì)出20組數(shù)據(jù)
            .size(20)
            // 添加文檔數(shù)據(jù)跟蹤
            .subAggregation(topHitsBuilder);
    searchSourceBuilder.aggregation(termsBuilder);
    request.source(searchSourceBuilder);
    SearchResponse response = client.search(request, RequestOptions.DEFAULT);
    // 取出統(tǒng)計(jì)結(jié)果
    Terms terms = response.getAggregations().get("schoolGroup");
    for (Terms.Bucket bucket : terms.getBuckets()) {
        System.out.println(bucket.getKeyAsString() + " 的人數(shù)：" + bucket.getDocCount());
        // 取出topHits跟蹤的文檔數(shù)據(jù)
        TopHits groupData = bucket.getAggregations().get("groupData");
        for (SearchHit hit : groupData.getHits()) {
            System.out.println(hit.getSourceAsString());
        }
        System.out.println("---------------------------------------------------------------------------");
    }
}

聚合查詢的相關(guān)內(nèi)容就介紹這么多了，更多的可以查看官方文檔稍浆。