s3-ls 0.0.3

Fast S3 object listing

s3-ls

Fast S3 object listing

Introduction

Listing S3 directory containing millions of objects is very painful because it has to be sequential, but maximum number of objects per one request is limited to 1000.

# First 1k objects
response = s3.list_objects_v2(
    Bucket=bucket, Prefix=prefix, StartAfter=""
)
last_key = response["Contents"][-1]["Key"]

# Next 1k objects
response = s3.list_objects_v2(
    Bucket=bucket, Prefix=prefix, StartAfter=last_key
)
...

s3-ls parallelizes it by partitioning the prefix into multiple parts and listing them in parallel.

prefix = "path/to/dir/"

prefix_1 = "path/to/dir/a"
prefix_2 = "path/to/dir/b"
prefix_3 = "path/to/dir/c"
... (70 prefixes)

In this way, the listing process can be done 70x faster in the ideal case.

Quick Start

Installation

pip install s3-ls

Usage

s3-ls s3://bucket/prefix

from s3_ls import list_objects

for obj in list_objects(bucket, prefix, **s3_kwargs):
    key = obj["Key"]