Sign-off exports - Atoti Hub

Export	Description
Store	Fact-level data exported from the store fields contributing to the Cube. This export type handles native measures without post-processing. The current implementation queries the Cube for facts and selects the levels corresponding to store fields.
Aggregated	Export of aggregated native measures queried from the Cube. Aggregation is achieved through removing fields from a store query (e.g. removing the TradeID results in Book-level aggregated data).
Cube	Export of any configured post-processed measures in the Cube.
Summary	Export of aggregated data at book level intended to be re-imported into the main Cube.
BookParentChild	Export of the current parent-child structure.
FXRates	Export of the current FX rate table.

To ensure the later re-import of data, during the Cube-level export we make the distinction between data that can be re-aggregated (e.g. PnL data at book level for historical trend analysis of VaR, linear aggregation) and data that cannot be re-aggregated (e.g. VaR data at book level, since VaR is a non-linear measure):

Cube exports are not meant to be reloaded in the MR application for re-aggregation.
Deprecated Aggregated exports are exports of linear measures that are meant to be reloaded in the MR application for re-aggregation in the Summary Cube. With the introduction of DirectQuery, this method should be considered as deprecated.
The Summary export files can be reloaded in the main Cube with the specific summary topic (if the Summary cubes are disabled).

Summary Exports

Export/import of sensitivity feature

Taylor Var

You have to choose to either use the exported sensitivity and also export the market shift, or to directly export the Taylor PnL Vector to import it back to the Var/ES Cube.

PnL explain

You have to choose to either use the exported sensitivity and also export the market data, or to directly export the PnL Explain to import it back to the PnL Cube.

Limitations

VaR / Taylor VaR multi-jurisdiction

The Market Shift Store is not exported, but it contains the shifts of the FX exchange rates needed to change the base currency. To keep this feature, those shifts need to be loaded onto the Market Shift Store.

Primary key collision

Depending on the daily input data, the exported, aggregated data may have primary key collision. In this case, the discriminator field has to be set up as a primary key on the target base store, or removed from the export template and replaced by a constant.

Remove the discriminator

Here is an example of removing the risk class for VaR, on VaRSummaryExport.json:

Remove from queries/locations: RiskClass@Risk Classes@Risk=[*];
Replace the Risk Class column definition by:

      {
        "@type": "echoOutputColumn",
        "header": "Risk Class",
        "value": "N/A"
      },

Add a primary key

For example, to set the risk class for VaR as a primary key, just define the following bean:

	@Bean
	public DatastoreConfiguratorConsumer signOffCustomisations() {
		return configurator -> configurator.updateFields("VaR", StoreConstants.TRADE_PNL_STORE_NAME, List.of(new CustomField(StoreFieldConstants.RISK_CLASS).asKeyField()));
	}

​Summary Exports

​Export/import of sensitivity feature

​Taylor Var

​PnL explain

​Limitations

​VaR / Taylor VaR multi-jurisdiction

​Primary key collision

​Remove the discriminator

​Add a primary key