Hey everyone, I’m stuck with a problem in Google Cloud’s Vertex AI Agent Builder. I created a Data Store for my chatbot and uploaded a CSV file with Q&A data. Now I want to change the primary key to the ‘Question’ field, but it’s not working.
When I try to update the schema, I get this error:
Schema update doesn't support adding the key property annotation for schema with active documents. Key property mapping mismatch for field 'Question'.
I’ve tried modifying the schema settings after uploading the data, but no luck. Has anyone faced this issue before? Any ideas on how to change the primary key without messing up the existing data? I’m really hoping to use the ‘Question’ field as the unique identifier for each entry.
Any help or workarounds would be much appreciated. Thanks in advance!
I’ve dealt with this exact problem in Vertex AI Agent Builder. It’s frustrating, but there’s a workaround that saved me hours of headache. Instead of changing the primary key directly, I created a new column called ‘QuestionID’ and used a simple Python script to generate unique IDs based on the ‘Question’ field. I then set ‘QuestionID’ as the primary key.
One approach is to add a new ‘QuestionID’ column to your CSV and generate unique identifiers using a hash function on the question text. After that, upload the modified CSV to a new data store. This method preserves your original data and still gives you a unique identifier for each entry. Be sure to update any code that references the data store to look up entries using ‘QuestionID’ instead of ‘Question’.
I’ve encountered a similar issue with Vertex AI Agent Builder. Unfortunately, changing the primary key on an existing data store with active documents isn’t supported directly. One workaround I’ve used is to create a new data store with the desired schema, then export the data from the old store and import it into the new one. It’s a bit tedious, but it works.
Another option is to add a new field as the primary key, populate it with unique values, and use that instead of modifying the existing ‘Question’ field. This approach might be less disruptive if you have a large dataset.
If these solutions aren’t feasible, you might need to contact Google Cloud support for more advanced options. They might be able to suggest a method that doesn’t require recreating the entire data store.
hey there, had similar trouble. quick fix: export data, nuke old store, make new one with right key. bit messy but works. remember to backup! if u need help, dm me. good luck mate!
I’ve run into this issue before with Vertex AI Agent Builder. Changing the primary key on an existing data store is indeed problematic. A solution that worked for me was to create a new data store with the desired schema, then use the Google Cloud SDK to export the data from the old store and import it into the new one. It’s a bit time-consuming, but it preserves your data integrity.
If you’re comfortable with Python, you could also write a script to automate this process. It would involve using the Vertex AI API to create the new store, export the old data, transform it to match the new schema, and then import it. This approach is more efficient for larger datasets.
Remember to thoroughly test your chatbot after migrating the data to ensure everything works as expected with the new primary key structure.
hey mate, i had the same headache with vertex AI. what worked for me was exportin the data, deletin the old store, and makin a new one with the right key. bit of a pain, but it did the trick. just make sure to backup everything first. good luck!