Salesforce Certified Platform Data Architect free question 10 / 10 · Domain 6: Data Migration
You are migrating contacts from a legacy system with Bulk API 2.0. The CSV contains rows with empty required fields and rows with values that are not in the picklist. A team member reasons that error rows can be picked up afterward and loads the file as is, but the load is much slower than expected. Which statement matches the large data volume best practices?
- Salesforce corrects the values before saving the rows, so data quality can be handled after the load instead of before it
- Error rows are dropped immediately and the remaining rows go through at full speed, so only the row count explains the slowness
- An error in a batch switches that batch to single-row processing, which hurts performance, so clean the data before loading
- One error causes the whole job to restart from the beginning, so split the file into smaller loads to cut down the rows
Show the answer and explanation
Answer: C
Why this is correct
The large data volume best practices (Loading Data from the API) say to make sure data is clean before loading when using Bulk API 2.0, because errors in batches trigger single-row processing for that batch, and that processing heavily impacts performance. Leaving error rows to be picked up later therefore costs speed as well as quality.
Why the other options are wrong
"Error rows are dropped immediately and the remaining rows go through at full speed, so only the row count explains the slowness" … The guide says an error in a batch triggers single-row processing for that batch, which heavily impacts performance. Error rows do slow the load; they are not irrelevant to speed.
"Salesforce corrects the values before saving the rows, so data quality can be handled after the load instead of before it" … The guide tells you to make sure the data is clean before loading with Bulk API 2.0. It does not describe Salesforce correcting values before saving them.
"One error causes the whole job to restart from the beginning, so split the file into smaller loads to cut down the rows" … What the guide describes is that a batch with errors is processed row by row. It does not say that the whole job restarts, and splitting the file would not help while the error rows remain.
Key points
Data quality at load time is about speed as well as correctness. Check required fields, picklist values, and data types before loading so that batches do not fall back to single-row processing.