Spring Boot Elasticsearch：RangeQuery中gte、lte、field引用异常

使用Spring Boot 3.4.1、Elasticsearch 8.15.5及Kotlin版Spring Data Elasticsearch构建日期范围查询时，RangeQuery.Builder的field、gte、lte方法出现未解析引用错误，其余过滤逻辑正常，仅日期范围过滤失效。

错误代码片段 #

val rangeQuery = RangeQuery.Builder()
    .field("lastSeen") // ❌ 未解析引用
    .gte(JsonData.of(startDate))  // ❌ 未解析引用
    .lte(JsonData.of(endDate))  // ❌ 未解析引用
    .build()

完整过滤函数 #

@Repository
class CustomCrashSummaryRepository(
    private val elasticsearchClient: ElasticsearchClient,
    private val elasticsearchOps: ElasticsearchOperations
) {
    fun findCrashSummariesByFilters(
        merchantId: String,
        projectId: String,
        osName: String,
        isIssueClosed: Boolean?,
        deviceModelList: List<String>?,
        osVersionList: List<String>?,
        appVersionList: List<String>?,
        pageable: Pageable,
        startDate: String?,
        endDate: String?
    ): Page<CrashSummary> {
        val query: NativeQuery = NativeQuery.builder()
            .withQuery { q ->
                q.bool { b ->
                    b.filter { f -> f.term { t -> t.field("merchantId.keyword").value(merchantId) } }
                    b.filter { f -> f.term { t -> t.field("projectId.keyword").value(projectId) } }
                    b.filter { f -> f.term { t -> t.field("osName.keyword").value(osName) } }
                    if (isIssueClosed != null) {
                        b.filter { f -> f.term { t -> t.field("isIssueClosed").value(isIssueClosed) } }
                    }
                    if (!deviceModelList.isNullOrEmpty()) {
                        b.filter { f ->
                            f.terms { t ->
                                t.field("devices.keyword")
                                    .terms { terms -> terms.value(deviceModelList.map { FieldValue.of(it) }) }
                            }
                        }
                    }
                    if (!osVersionList.isNullOrEmpty()) {
                        b.filter { f ->
                            f.terms { t ->
                                t.field("osVersions.keyword")
                                    .terms { terms -> terms.value(osVersionList.map { FieldValue.of(it) }) }
                            }
                        }
                    }
                    if (!appVersionList.isNullOrEmpty()) {
                        b.filter { f ->
                            f.terms { t ->
                                t.field("appVersions.keyword")
                                    .terms { terms -> terms.value(appVersionList.map { FieldValue.of(it) }) }
                            }
                        }
                    }

                    if (startDate != null || endDate != null) {

                        val rangeQuery = RangeQuery.Builder()
                            .field("lastSeen")
                            .gte(JsonData.of(startDate))
                            .lte(JsonData.of(endDate))
                            .build()

                        b.filter(rangeQuery._toQuery())

                    }
                    b
                }
            }
            .withPageable(pageable)
            .build()

        val searchResponse = elasticsearchOps.search(query, CrashSummary::class.java)
        val contentList: List<CrashSummary> = searchResponse.map { it.content }.toList()
        val totalHits = searchResponse.totalHits
        return PageImpl(contentList, pageable, totalHits)
    }
}

导入语句 #

import co.elastic.clients.elasticsearch.ElasticsearchClient
import co.elastic.clients.elasticsearch.core.UpdateByQueryRequest
import co.elastic.clients.elasticsearch._types.Script
import co.elastic.clients.elasticsearch._types.FieldValue
import co.elastic.clients.elasticsearch._types.aggregations.TermsAggregation
import co.elastic.clients.elasticsearch._types.query_dsl.*
import co.elastic.clients.elasticsearch.core.SearchRequest
import co.elastic.clients.json.JsonData
import co.elastic.clients.json.jackson.JacksonJsonpMapper
import com.fasterxml.jackson.databind.ObjectMapper
import io.initial.logdrop.core.entity.CrashSummary
import io.initial.logdrop.core.entity.admin.CrashFilterResult
import org.springframework.stereotype.Repository
import java.io.IOException
import org.springframework.data.domain.PageImpl
import java.io.StringWriter
import org.springframework.data.domain.Page
import org.springframework.data.domain.Pageable
import org.springframework.data.elasticsearch.client.elc.NativeQuery
import org.springframework.data.elasticsearch.core.ElasticsearchOperations

文档映射（lastSeen字段） #

...
  "lastSeen": {
    "type": "date",
    "format": "epoch_millis"
  }
...

示例数据 #

{
  "description": [
    "java.lang.NullPointerException"
  ],
  "firstSeen": [
    "2025-03-17T16:12:46.887Z"
  ],
  "isIssueClosed": [
    false
  ],
  "issueId": [
    "1651584580"
  ],
  "issueId.keyword": [
    "1651584580"
  ],
  "lastSeen": [
    "2025-03-17T16:12:46.887Z"
  ]
}

问题原因 #

Elasticsearch 8.x Java客户端中，RangeQuery.Builder采用嵌套配置设计，无法直接调用.field()、.gte()等方法，必须通过.range()配置块才能访问这些方法。

修复步骤 #

1. 调整RangeQuery.Builder的构建逻辑，用.range()配置块包裹字段和范围条件
2. 处理startDate或endDate为null的场景，避免传入无效的JsonData

修改后的日期范围查询代码 #

if (startDate != null || endDate != null) {
    val rangeQuery = RangeQuery.Builder()
        .range { r ->
            var rangeBuilder = r.field("lastSeen")
            startDate?.let { rangeBuilder = rangeBuilder.gte(JsonData.of(it)) }
            endDate?.let { rangeBuilder = rangeBuilder.lte(JsonData.of(it)) }
            rangeBuilder
        }
        .build()

    b.filter(rangeQuery._toQuery())
}

额外优化（可选） #

由于lastSeen字段映射为epoch_millis，如果传入的日期是ISO字符串格式（如示例中的2025-03-17T16:12:46.887Z），建议先转换为时间戳再传入，避免Elasticsearch自动解析可能出现的格式问题：

import java.time.Instant

// 日期转换示例
startDate?.let {
    val epochMillis = Instant.parse(it).toEpochMilli()
    rangeBuilder = rangeBuilder.gte(JsonData.of(epochMillis))
}

内容的提问来源于stack exchange，提问作者Ozgur Baykal