Error: DynamoDB batch write is not working

A batch write is incomplete if it returns UnprocessedItems, the call was never awaited, or buffered writes were not flushed. A successful HTTP response alone does not prove every item was written. Check both request validation and per-batch progress.

Boto3’s Table.batch_writer handles buffering and resubmits unprocessed items. Exit its context normally and let errors propagate:

import boto3

table = boto3.resource("dynamodb", region_name="us-east-1").Table("Orders")
with table.batch_writer() as batch:
    batch.put_item(Item={"id": "123", "status": "NEW"})
    batch.put_item(Item={"id": "456", "status": "NEW"})

For low-level batch_write_item or JavaScript BatchWriteCommand, implement the UnprocessedItems loop explicitly with a bounded recovery policy. Send at most 25 puts or deletes per request, match the complete key, and remove duplicate operations on the same item within a batch.

Batch puts replace existing items rather than merging fields. BatchWriteItem cannot perform updates or conditional writes and is not atomic as a whole. Diagnose persistent throttling on the actual table or GSI; changing SDK method names does not eliminate capacity limits.

See bounded batch retry example and AWS documentation.