DynamoDB Scan in Python (boto3)

scan reads every item in a table or index, page by page, then optionally drops rows with a filter. With the boto3 resource API you express the filter with the Attr helper from boto3.dynamodb.conditions.

Code

import boto3
from boto3.dynamodb.conditions import Attr

dynamodb = boto3.resource("dynamodb", region_name="us-east-1")
table = dynamodb.Table("Music")

def scan_recent_songs():
    items = []
    kwargs = {"FilterExpression": Attr("Year").gte(2010)}

    while True:
        response = table.scan(**kwargs)
        items.extend(response["Items"])

        # Keep going until every 1 MB page is read
        if "LastEvaluatedKey" not in response:
            break
        kwargs["ExclusiveStartKey"] = response["LastEvaluatedKey"]

    print(f"Matched {len(items)} songs")
    return items

if __name__ == "__main__":
    scan_recent_songs()

Explanation

• Attr("Year").gte(2010) — the Attr helper builds the filter expression, names, and values. Combine conditions with & / | (e.g. Attr("Genre").eq("Jazz") & Attr("Year").gte(2010)). Year being a reserved word is handled for you.
• Filtering is post-read — it shrinks the response, not the cost. You pay to read every item scanned.
• Pagination is mandatory — each page reads at most 1 MB before filtering, so a filtered page can be empty while LastEvaluatedKey is still set. Loop until it's absent or you'll miss matches.
• For a large full-table scan, parallelize with Segment / TotalSegments.
• Numeric attributes return as decimal.Decimal — compare against Decimal/int, not float.
• Prefer query. scan reads the whole table; on anything but a tiny table it's slow and expensive.

Do it visually

The DynamoDB Expression Builder assembles the Attr(...) filter and copies runnable code.

To explore tables in a GUI — filtered, paginated result grids and copy-as-code — download DynoTable, instead of scanning blind from a script.

Related guides

• Query vs. Scan — when (rarely) a scan is justified.
• Why is my DynamoDB Scan slow and expensive? — the cost model and how to avoid it.