The download_file method is the simplest way to save an S3 object directly to your local file system. You provide the bucket name, the object key, and the specific path on your computer where the file should be saved.
import boto3 s3 = boto3.client('s3') s3.download_file('my-bucket', 'data/report.pdf', '/tmp/report.pdf') Use code with caution. s3 download_file vs download_fileobj
Both methods are "managed transfers," meaning they automatically handle multipart downloads and multithreading for large files to maximize performance. The download_file method is the simplest way to
Flexibility. You can use a BytesIO buffer to keep data in memory or an already-open file handle. Usage Example: Both methods are "managed transfers
The download_fileobj method accepts a "file-like object" that has already been opened in . This is ideal when you don't want to save the data to a disk—for instance, if you are processing it immediately in a Lambda function or streaming it to another service.
When downloading objects from Amazon S3 using Boto3 , the choice between download_file and download_fileobj depends primarily on where you want the data to land: a named file on your hard drive or an open "file-like" object in memory.
import boto3 import io s3 = boto3.client('s3') buffer = io.BytesIO() # Download into the buffer s3.download_fileobj('my-bucket', 'data/report.pdf', buffer) # Rewind the buffer to read from the start buffer.seek(0) data = buffer.read() Use code with caution. download_file download_fileobj Target Destination A local file path (string) A writeable file-like object Managed Transfer Yes (multipart/threaded) Yes (multipart/threaded) Best Use Case Saving to a hard drive Buffering in memory or streaming Mode Must be binary mode ( wb ) Performance and Memory Considerations Stack Overflow