Before running your first migration, ensure you have a supported Java version installed (JDK 11 or later). Next, download the latest ora2iceberg binary from the official GitHub releases page, or build it from source if preferred. Once the executable is ready, test it by running ora2iceberg without parameters to confirm the tool’s basic functionality.
After confirming the setup, take time to examine the Oracle table(s) you plan to migrate. Review the schema, focusing on columns defined with the NUMBER type without explicit precision and scale. Consider the intended semantic meaning of each column:
IDs and Keys: Often best represented as long in Iceberg to ensure efficient indexing and lookups.
Monetary Values: May require a fixed decimal scale (e.g., decimal(18,2)) to accurately capture cents or subunits.
Large Integers or High-Precision Values: Choose a suitable decimal type (e.g., decimal(38,10)) for columns holding large or complex numeric data.
This analysis will guide you in using parameters like --default-number-type and --data-type-map to ensure that your numeric data is correctly and consistently mapped into Iceberg. Once you understand the source structure, you’ll be better positioned to configure ora2iceberg for a successful and meaningful data migration.