When you enable DynamoDB global tables, the table you're creating replicas for must be empty. This is because global tables use a process called "table copying" to replicate data across regions. The table copying process only replicates the data present when the replicas are created.
When you create global table replicas, the primary replica table is used as the source table, and the data in the table is replicated to the other regions. However, if the primary replica table already contains data, it would be difficult to determine which data should be replicated and which data should not be replicated. Therefore, creating the primary replica table as an empty table is necessary.
Another reason is that DynamoDB global tables use DynamoDB Streams to propagate the data changes to the replica tables. If the primary replica table already contains data, it would be difficult to determine which data changes should be propagated and which data changes should not be propagated. Therefore, creating the primary replica table as an empty table is necessary.
It's important to note that once the global tables are enabled, you can insert, update or delete data in the primary table. The data changes would be propagated to the replica tables using DynamoDB Streams, ensuring that all the replica tables are up-to-date.
Other Common DynamoDB FAQ (with Answers)
- Why is DynamoDB easy to partition?
- Can DynamoDB have null values?
- How to connect to DynamoDB?
- Is DynamoDB based on MongoDB?
- Are DynamoDB tables globally unique?
- What is DynamoDB white paper, and what are the key takeaways?
- Can QuickSight read DynamoDB?
- Is DynamoDB serverless?
- Is DynamoDB columnar database?
- Is DynamoDB multi-region?
- Can DynamoDB be replicated?
- What is DynamoDB used for?
- Is DynamoDB a memory store?
- Is DynamoDB a key-value store?
- Does DynamoDB Support SQL?