Java代码解决ElasticSearch的Result window is too large问题

调用elasticsearch做分页查询时报错：

queryphaseexecutionexception[result window is too large, from + size must be less than or equal to: [10000] but was [666000]. see the scroll api for a more efficient way to request large data sets. this limit can be set by changing the [index.max_result_window] index level setting.]; }

提示用from+size方式有1万条数据查询的限制，需要更改index.max_result_window参数的值。

翻了下elasticsearch官网的文档：

index.max_result_window
the maximum value of from + size for searches to this index.defaults to 10000. 
search requests take heap memory and time proportional to from + size and this limits that memory. 
see scroll or search after for a more efficient alternative to raising this.

说是用传统方式（from + size）查询占用内存空间且比较消耗时间，所以做了限制。

问题是用scroll方式做后台分页根本行不通。

不说用scroll方式只能一页页的翻这种不人性化的操作。页码一多，scrollid也很难管理啊。

所以继续鼓捣传统方式的分页。

上网查了下设置max_result_window的方法，全都是用crul或者http方式改的。

后来无意间看到了一篇文档： https://blog.csdn.net/tzconn/article/details/83309516

结合之前逛elastic中文社区的时候知道这个参数是索引级别的。于是小试了一下，结果竟然可以了。

java代码如下：

public searchresponse search(string logindex, string logtype, querybuilder query, 
                 list<aggregationbuilder> agg, int page, int size) {
        page = page > 0 ? page - 1 : page;
        transportclient client = getclient();
        searchrequestbuilder searchrequestbuilder = client.preparesearch(logindex.split(","))
                .settypes(logtype.split(","))
                .setsearchtype(searchtype.dfs_query_then_fetch)
                .addsort("createtime", sortorder.desc);

        if (agg != null && !agg.isempty()) {
            for (int i = 0; i < agg.size(); i++) {
                searchrequestbuilder.addaggregation(agg.get(i));
            }
        }
        updateindexs(client, logindex, page, size);

        searchresponse searchresponse = searchrequestbuilder
                .setquery(query)
                .setfrom(page * size)
                .setsize(size)
                .get();
        return searchresponse;
    }

    //更新索引的max_result_window参数
    private boolean updateindexs(transportclient client, string indices, int from, int size) {
        int records = from * size + size;
        if (records <= 10000) return true;
        updatesettingsresponse indexresponse = client.admin().indices()
                .prepareupdatesettings(indices)
                .setsettings(settings.builder()
                        .put("index.max_result_window", records)
                        .build()
                ).get();
        return indexresponse.isacknowledged();
    }

搞定。

当然这段代码不好的地方在于：

每次查询超过10000万条记录的时候，都会去更新一次index。

这对原本就偏慢的from+size查询来说，更是雪上加霜了。