Problem
Ingest benchmarks using insertBatch can only be run once against a given data file, since the documents are inserted with their original IDs. On subsequent runs, the inserts fail with duplicate key errors.
This makes it difficult to iterate on ingest benchmarks or re-run them without manually clearing the database or preparing a new data file each time.
A potential option could be to automatically generate unique IDs for each run (e.g. prefix or suffix the document ID with a run identifier).
Problem
Ingest benchmarks using
insertBatchcan only be run once against a given data file, since the documents are inserted with their original IDs. On subsequent runs, the inserts fail with duplicate key errors.This makes it difficult to iterate on ingest benchmarks or re-run them without manually clearing the database or preparing a new data file each time.
A potential option could be to automatically generate unique IDs for each run (e.g. prefix or suffix the document ID with a run identifier).