What is it?

When data is propagated to data sources to the UDM, our entity resolution process works behind the scenes to identify and link identical records from multiple different sources. Once the specific set of entity resolution rules are configured, this process happens automatically at ingestion time—the user cannot influence the result.

For most use cases, this approach works well. However, because there is limited visibility on the actions taken during ingestion time, it can be difficult to anticipate exactly how a record will be interpreted by the system. This tool offers a way to preview how entity resolution will behave before data is submitted — helping you better understand and test the matching logic. While it can also support specific workflows such as preventing duplicate records or improving data quality, its primary purpose is to provide visibility into the resolution process.

How does it work?

Before we can dive into how Entity Resolution Preview works, let’s get familiar with the configuration we’ll use as an example.

• We have 2 data sources, A and B, that represents 2 different systems we are sourcing data from
• Each data source contains clinic resources that contains a name and an address.
• We have an MDM collection to create a unified record of type clinic that has resolvers with matcher rules for an exact name and a partial address (same postal code, city, country and state).

Let’s say that there exists an MDM record of type clinic from Data Source A. We’d like to find out what would happen if we were add a very similar clinic into Data Source B. In the following data, we can see that addresses are exactly the same but the name differs, which should give us a “possible match” outcome if we were to ingest it.

Using the Entity Resolution Preview feature, you can see the outcome before actually ingesting the new record. In this case, the data from the new clinic that is about to be ingested into Data Source B will compared to the MDM to find any existing records. Since there already exists a unified clinic record with a similar address, this will be flagged as a “possible match” due to the address having a match on the postal code, city, state and country.

When calling the Entity Resolution Preview endpoint with the data you want to ingest into Data Source B, you will receive a response with a list of unified records that the system has matched using the configured resolution rules with their resolution outcome. This will give you insight on how the system will handle reconciling your incoming data and allow you to make an informed decision.