cancel
Showing results for 
Search instead for 
Did you mean: 
Community Platform Discussions
Connect with fellow community members to discuss general topics related to the Databricks platform, industry trends, and best practices. Share experiences, ask questions, and foster collaboration within the community.
cancel
Showing results for 
Search instead for 
Did you mean: 

Using DeltaTable.merge() and generating surrogate keys on insert?

Dekova
New Contributor II

I'm using merge to upsert data into a table:

DeltaTable
.forName(DESTINATION_TABLE)
.as("target")
.merge(merge_df.as("source") ,"source.topic = target.topic and source.key = target.key")
.whenMatched()
.updateAll()
.whenNotMatched()
.insertAll()
.execute()

Id like to use one of the following generated columns to create surrogate keys on insert:

  1. surrogate_guid string generated always as (uuid())
  2. surrogate_id bigint generated always as identity

#1 doesn't work, I get the error "A generated column cannot use a non-deterministic expression".
#2 doesn't work, I get the error "cannot resolve surrogate_id in UPDATE clause"

What's best practice for merging and getting a unique identifier (preferably uuid) assigned?

1 ACCEPTED SOLUTION

Accepted Solutions

daniel_sahal
Esteemed Contributor

@Dekova 
1) uuid() is non-deterministic meaning that it will give you different result each time you run this function
2) Per the documentation "For Databricks Runtime 9.1 and above, MERGE operations support generated columns when you set spark.databricks.delta.schema.autoMerge.enabled to true."

What i would do in this situtaion is:
- Create a column surrogate_id bigint GENERATED ALWAYS AS identity,
- Create a column surrogate_guid and hash it based on surrogate_id column. Ex. surrogate_guid string GENERATED ALWAYS AS (sha2(cast(surrogate_id AS STRING), 512))

View solution in original post

1 REPLY 1

daniel_sahal
Esteemed Contributor

@Dekova 
1) uuid() is non-deterministic meaning that it will give you different result each time you run this function
2) Per the documentation "For Databricks Runtime 9.1 and above, MERGE operations support generated columns when you set spark.databricks.delta.schema.autoMerge.enabled to true."

What i would do in this situtaion is:
- Create a column surrogate_id bigint GENERATED ALWAYS AS identity,
- Create a column surrogate_guid and hash it based on surrogate_id column. Ex. surrogate_guid string GENERATED ALWAYS AS (sha2(cast(surrogate_id AS STRING), 512))

Connect with Databricks Users in Your Area

Join a Regional User Group to connect with local Databricks users. Events will be happening in your city, and you won’t want to miss the chance to attend and share knowledge.

If there isn’t a group near you, start one and help create a community that brings people together.

Request a New Group