Skip to content

Data Ingestion

Hongzheng Shi edited this page Dec 3, 2018 · 2 revisions

Data ingestions are achieved via upsert batches with the following procedures:

  1. Client contacts the server for schema resolution before preparing an upsert batches. This converts column names to column IDs.
  2. Client contacts the server for auto-enum value resolution. This converts all string values of auto-enum columns to integers. In distributed setup all servers need to agree on the same schema and auto-enum dictionary.
  3. Client prepares an upsert batch using the serialized format. Unlike a SQL insert, AresDB does not provide transaction atomicity and isolation higher than record level.
You can’t perform that action at this time.