Multiple Data Services

A domain service can work with several data services for the same entity - typically a database and a Redis cache in front of it. Instead of writing the "read from the cache, fall back to the database, then cache the result" logic yourself, you describe it with three options on any domain method call.

// read from the cache first; if it's not there, read from the database and save it in the cache
const { result } = await coursesService.findOne({
  filters: { id: 5 },
  dataServices: ['cache', 'main'],
  saveAdditionalResultsInFirstService: { serviceName: 'main', useResultsForFirstServiceReturnData: true }
});

The data services a domain service can use are the ones passed to its constructor (see Domain services): main is the first constructor argument, and the others are named in the last one, e.g. { cache: DataCacheCoursesEntityService }.

The options

• dataServices?: string[] - the services to use, in order. Defaults to ['main']. The first one is the first service, and its result is returned as result.
• optionsOverridesByService?: { [serviceName]: options } - options for the additional services. Override filters are combined with the original ones using $and, and the other options replace the original ones. Also supports:
  • runOnNoFirstServiceResultOnly - defaults to true: an additional service only runs if the first one returned nothing. Set it to false to always run it.
  • filterByFirstServiceResultFields - filter the additional service by fields of the first service's result, as { sourceField: targetField } (e.g. { id: 'id' }). It's skipped when the first service returned nothing, so it's normally combined with runOnNoFirstServiceResultOnly: false.
  • returnData, returnOriginalItems
• saveAdditionalResultsInFirstService?: { serviceName, useResultsForFirstServiceReturnData?, saveOptions? } (find and findOne only) - save what the named additional service found into the first service. With useResultsForFirstServiceReturnData, those results are returned as result when the first service found nothing.

The response is { result, resultsByService? } - resultsByService has each additional service's result, by name.

How each method uses the additional services

Method	Additional services get	They run by default when
find, findOne	the same options, plus their overrides	the first service found nothing
create, bulkCreate	the first service's result as the data (e.g. with the database-generated id)	never - a successful create is a result, so pass runOnNoFirstServiceResultOnly: false
update, delete	the same data and options	the first service changed something (count > 0)

WARNING

create with dataServices: ['main', 'cache'] does not write to the cache unless you also pass optionsOverridesByService: { cache: { runOnNoFirstServiceResultOnly: false } }.

TIP

Additional services get a deep copy of the options, so changes the first service makes to them don't leak into the others.

All three options can also come from the REST API's request body or query string (see Controllers). For a complete example with a Redis cache, see the Caching recipe.

---

Previous: Logging · Next: HTTP API module