Design and Implementation of a Configurable OLAP Framework for Rapid Development

Background: In many projects, building OLAP systems with multi‑dimensional, multi‑metric analysis is complex, time‑consuming, and data‑intensive, requiring significant development effort.

Architecture Design: The proposed framework separates a core OLAP domain from a front‑end interaction domain, using configurable metadata and a façade service to translate requests into SQL for column‑store databases, maximizing query performance.

Detailed Design – Runtime: The façade service distinguishes request types (table, chart, export), passes the OLAP domain model to the domain service, which generates executable SQL and returns results in a unified format.

Core Domain Model: Defines the four basic OLAP operations (roll‑up, drill‑down, slice, dice) and abstracts aggregation SQL into dimensions, measures, filters, grouping, ordering, and pagination.

Metadata Configuration: Provides configurable metric metadata and module metadata for tables, charts, and export, illustrated with JSON examples.

{
"cbj_amount"
: {
"expression"
:
"sum(cbj_amount)"
,
"name"
:
"cbj_amount"
,
"desc"
:
"损失分布-仓报价金额"
},
"sale_qtty"
: {
"expression"
:
"sum(sale_qtty)"
,
"name"
:
"sale_qtty"
,
"desc"
:
"销量"
},
"loss_amount"
: {
"expression"
:
"sum(loss_amount)"
,
"name"
:
"loss_amount"
,
"desc"
:
"损额"
}
}

Table configuration example:

{
"list-overall-lossCategory"
: {
"moduleName"
:
"整体诊断结果-损失类别"
,
"tableName"
:
"logis_loss_reason_nume_denom_month_dis"
,
"measureList"
: [
"dakuru_loss_ratio"
,
"dakuru_loss_ratio_yoy_chg"
,
"dakuru_loss_ratio_mom_chg"
,
"peisongru_loss_ratio"
,
"peisongru_loss_ratio_yoy_chg"
,
"peisongru_loss_ratio_mom_chg"
,
"shouhouru_loss_ratio"
,
"shouhouru_loss_ratio_yoy_chg"
,
"shouhouru_loss_ratio_mom_chg"
,
"dakupyk_loss_ratio"
,
"dakupyk_loss_ratio_yoy_chg"
,
"dakupyk_loss_ratio_mom_chg"
,
"bjkpyk_loss_ratio"
,
"bjkpyk_loss_ratio_yoy_chg"
,
"bjkpyk_loss_ratio_mom_chg"
],
"displayType"
:
"table"
,
"page"
: {
"pageNo"
:
1
,
"pageSize"
:
1
}
}
}

Chart configuration example:

{
"chart-lossType-trend-lossRatio"
: {
"moduleName"
:
"损失类别走势-损率"
,
"tableName"
:
"logis_loss_reason_nume_denom_month_dis"
,
"dimList"
: [
"mt"
],
"measureList"
: [
"dakuru_loss_ratio"
,
"peisongru_loss_ratio"
,
"shouhouru_loss_ratio"
,
"bjkpyk_loss_ratio"
,
"dakupyk_loss_ratio"
],
"displayType"
:
"chart"
,
"page"
: {
"pageNo"
:
1
,
"pageSize"
:
13
},
"chart"
: {
"categoryName"
:
"损失类别走势"
,
"categoryKey"
:
"mt"
,
"seriesList"
: [
{
"key"
:
"大库入"
,
"value"
:
"dakuru_loss_ratio"
},
{
"key"
:
"配送入"
,
"value"
:
"peisongru_loss_ratio"
},
{
"key"
:
"售后入"
,
"value"
:
"shouhouru_loss_ratio"
},
{
"key"
:
"备件库盘盈亏"
,
"value"
:
"bjkpyk_loss_ratio"
},
{
"key"
:
"大库盘盈亏"
,
"value"
:
"dakupyk_loss_ratio"
}
]
}
}
}

Export configuration example:

{
"export-lossType-trend"
: {
"moduleName"
:
"损失类别-损失类别走势-导出"
,
"tableName"
:
"logis_loss_reason_nume_denom_month_dis"
,
"dimList"
: [
"mt"
,
"loss_reason1"
],
"measureList"
: [
"distribute_loss_ratio"
,
"loss_amount"
,
"loss_qtty"
],
"filterList"
: [{
"lossReason1"
: {
"expr"
:
"is not null"
,
"value"
:
""
} }],
"headList"
: [
"mt"
,
"loss_reason1"
,
"distribute_loss_ratio"
,
"loss_amount"
,
"loss_qtty"
],
"fileName"
:
"损失类别走势信息"
,
"displayType"
:
"file"
,
"page"
: {
"pageNo"
:
1
,
"pageSize"
:
1000
,
"orderList"
: [ {
"orderField"
:
"mt"
,
"orderType"
:
"DESC"
}, {
"orderField"
:
"loss_reason1"
,
"orderType"
:
"DESC"
} ] }
}
}

System Integration: By adding the Maven dependency, configuration files, and implementing the required extension points, developers can quickly integrate the framework into existing back‑end services.

<dependency>
<groupId>com.jd.netsim</groupId>
<artifactId>triage-framework-core</artifactId>
<version>1.0-SNAPSHOT</version>
</dependency>

Conclusion: The article presents a generic, configurable OLAP framework that accelerates development of analytical applications, with potential for visual configuration and self‑service analytics, while acknowledging some details are omitted for brevity.