Add lookups from data sources
This page provides a guide on how to add lookups from data sources in Upsolver.
Last updated
This page provides a guide on how to add lookups from data sources in Upsolver.
Last updated
1. Click Add Lookup or in the Calculated Fields tab, under Lookups, click.
2. In the Data Source tab, select a source for your lookup.
3. Click Next.
4. Under Join Fields, select the fields to use as key columns.
5. Select at least one aggregation and the field to apply it to. See: Aggregation functions
6. Select the desired time frame; this enables you to sync your data sources.
When joining two streams, they may not be completely in sync (there may be a lead or lag).
An event in data source A may arrive before or after the corresponding event in data source B. The time frame is the period before and after the event in data source A arrives, to look for the join event in data source B.
Example: If time frame is set to -1 and 2 hours, the system looks for a matching event in data source B in the period one hour before and two hours after the time the event arrived in data source A.
You can also set the time frame to infinite by checking Infinite From. This ensures you can access all the data irrespective of when it was written.
7. Check Use Full History to use all the data that was available when deploying this output for historical lookups.
When joining a stream and a lookup, generally you use the time field for the join. As you move forward with the stream, you join to the lookup based on the time in the lookup.
This applies to the past data too; if processing an output for the past 12 months, the time in the lookup is used to match to the stream. This will not work in cases such as joining a stream with a year of data to a static table which has just recently been dumped.
In this case in order to join this data, check Use Full History. The time of the original event is then disregarded; instead the current time is used to perform the join (e.g. it uses the latest state of the data).
8. Decide whether or not to implement an inner join.
9. Select where to locate the calculated field (e.g. root, data) and name this field. This operation joins the input arrays using a cartesian product.
The fields are typically added at the root level, however on occasion you can opt to co‑locate a field with another element in the hierarchy.
10. Click Preview to review your calculated field before saving; a Field Content Samples panel will appear on the right.
11. To add columns to the preview, click Choose Columns and select the desired columns. Then click Update.
12. Click Save.
You can now find this lookup under Lookups in the sidebar or under the Calculated Fields tab.