Java与Elasticsearch的API实现工具类（二）——增删改，以及批量插入_java elasticsearch修改数据使用immediate

接着上面的项目，添加工具类的方法

项目gitee https://gitee.com/gangye/elasticsearch_demo

一、向es的指定索引中新增文档

    /**
     * @Description 单条记录插入
     * @param indexName
     * @param sourceMap  传入的记录的键值对
     * @return
     */
    public static Map<String, Object> singleInsert (String indexName, Map<String, Object> sourceMap) {
        Map<String, Object> insertResult = new HashMap<>();
        RestHighLevelClient client = EsClient.getInstance();
        IndexRequest indexRequest = new IndexRequest(indexName, DefineConstant.SEARCH_REQUEST_TYPE);
        //es对应的索引文档的字段一定要有id_s，便于与_id相对应，且一定要逻辑控制id的唯一
        String id = StringUtils.isEmpty(sourceMap.get("id_s")) ? UUID.randomUUID().toString() : (String) sourceMap.get("id_s");
        indexRequest.id(id);
        indexRequest.source(sourceMap);
        //写入后立即刷新
        indexRequest.setRefreshPolicy(WriteRequest.RefreshPolicy.IMMEDIATE);
        try {
            String lineSeparator = System.lineSeparator();
            log.info(lineSeparator + "index:" + indexName + lineSeparator + "insert:"+ indexRequest.toString() + lineSeparator);
            IndexResponse indexResponse = client.index(indexRequest, RequestOptions.DEFAULT);
            log.info("es insert response: {}", indexResponse.toString());
            insertResult.put("id", indexResponse.getId());
            insertResult.put("status", indexResponse.status().toString());
        } catch (IOException e) {
            log.error(e.getMessage());
        }
        return insertResult;
    }

 向income_expense_analysis索引中插入一条记录

对应的响应报文以及日志打印

 成功插入一条记录，这里的id_s与_id一致，便于后续修改操作的时候传入id 

 二、向es指定的索引中，根据文档id修改记录

    /**
     * @Description 修改记录数据
     * @param indexName
     * @param updateMap
     * @return
     */
    public static Map<String, Object> updateEsRecord (String indexName, Map<String, Object> updateMap){
        Map<String, Object> updateResult = new HashMap<>();
        RestHighLevelClient client = EsClient.getInstance();
        //此时之前指定的id_s字段就起作用了
        if (StringUtils.isEmpty(updateMap.get("id_s"))){
            log.error("尚未指定修改的记录id");
            return null;
        }
        String id = (String) updateMap.get("id_s");
        UpdateRequest updateRequest = new UpdateRequest(indexName, DefineConstant.SEARCH_REQUEST_TYPE, id);
 
        updateRequest.doc(updateMap);
        updateRequest.setRefreshPolicy(WriteRequest.RefreshPolicy.IMMEDIATE);
        try {
            String lineSeparator = System.lineSeparator();
            log.info(lineSeparator + "index:" + indexName + lineSeparator + "update:"+ updateRequest.toString() + lineSeparator);
            UpdateResponse updateResponse = client.update(updateRequest, RequestOptions.DEFAULT);
            log.info("es update response: {}", updateResponse.toString());
            updateResult.put("id", updateResponse.getId());
            updateResult.put("status", updateResponse.status().toString());
        } catch (IOException e) {
            log.error(e.getMessage());
        }
        return updateResult;
    }

 修改刚新增id为2020-03-313186482的文档的txn_remark_s字段为"03月31日商户交易入账"

 对应的响应报文以及日志打印

成功修改文档记录

三、根据指定的id删除文档

    /**
     * @Description 删除记录数据
     * @param indexName
     * @param id
     * @return
     */
    public static Map<String, Object> deleteEsRecord (String indexName, String id){
        Map<String, Object> deleteResult = new HashMap<>();
        RestHighLevelClient client = EsClient.getInstance();
        //此时之前指定的id_s字段就起作用了
        if (StringUtils.isEmpty(id)){
            log.error("尚未指定删除的记录id");
            return null;
        }
        DeleteRequest deleteRequest = new DeleteRequest(indexName, DefineConstant.SEARCH_REQUEST_TYPE, id);
        deleteRequest.setRefreshPolicy(WriteRequest.RefreshPolicy.IMMEDIATE);
        try {
            String lineSeparator = System.lineSeparator();
            log.info(lineSeparator + "index:" + indexName + lineSeparator + "delete:"+ deleteRequest.toString() + lineSeparator);
            DeleteResponse deleteResponse = client.delete(deleteRequest, RequestOptions.DEFAULT);
            log.info("es update response: {}", deleteResponse.toString());
            deleteResult.put("id", deleteResponse.getId());
            deleteResult.put("status", deleteResponse.status().toString());
        } catch (IOException e) {
            log.error(e.getMessage());
        }
        return deleteResult;
    }

删除刚刚新增的id为2020-03-313186482 的文档记录

响应的结果以及日志打印

 再结合工具查看income_expense_analysis索引下的文档数据记录，成功删除

四、批量插入数据，使用bulk操作

    /**
     * @Description  批量向es插入数据，最好将批量数据分批向es插入，而且将客户端的超时时间延长(此处我使用的单记录的操作的5秒对于批量操作就不合适了)
     * @param indexName
     * @param sourceList
     * @return
     */
    public static Map<String, Object> batchInsert(String indexName, List<Map<String, Object>> sourceList){
        List<Map<String, Object>> insertResults = new ArrayList<>();
        Map<String, Object> result = new HashMap<>();
        RestHighLevelClient client = EsClient.getInstance();
        BulkRequest bulkRequest = new BulkRequest();
        for (Map<String, Object> source : sourceList){
            IndexRequest indexRequest = new IndexRequest(indexName,DefineConstant.SEARCH_REQUEST_TYPE);
            indexRequest.id(StringUtils.isEmpty(source.get("id_s")) ? UUID.randomUUID().toString() : (String) source.get("id_s"));
            indexRequest.source(source);
            bulkRequest.add(indexRequest);
        }
        bulkRequest.timeout(TimeValue.timeValueMintue(2));
        bulkRequest.setRefreshPolicy(WriteRequest.RefreshPolicy.WAIT_UNTIL);
        try {
            BulkResponse bulkResponse = client.bulk(bulkRequest, RequestOptions.DEFAULT);
            BulkItemResponse[] itemResponses = bulkResponse.getItems();
            for (int i = 0;i < itemResponses.length; i++){
                BulkItemResponse response = itemResponses[i];
                Map<String, Object> item = new HashMap<>();
                item.put("id", response.getId());
                item.put("status", response.status().toString());
                item.put("failure", response.getFailureMessage());
                if (response.isFailed()){
                    IndexRequest ireq = (IndexRequest) bulkRequest.requests().get(i);
                    log.error("Failed while indexing to " + response.getIndex() + "type " +response.getType() + " " +
                            "request: [" + ireq + "]: [" + response.getFailureMessage() + "]");
                    item.put("isSuccess", "FAIL");
                }
                log.info("-------bulk---batchInsert------->{}", item);
                insertResults.add(item);
            }
        } catch (IOException e) {
            log.error(e.getMessage());
        }
        result.put("insertResults", insertResults);
        return result;
    }

一般是从文件中读取记录，此处我就不使用请求中展示，就在代码中模拟插入的记录

 响应结果以及执行的日志打印

查看es插入的记录