# How to access blockchain data from AWS S3 - Parquet Format

**URL:** <https://community.bitquery.io/t/how-to-access-blockchain-data-from-aws-s3-parquet-format/1630>\
**Category:** GraphQL Tutorials\
**Tags:** support, cloud\
**Created:** [August 29, 2023, 12:27pm UTC](https://community.bitquery.io/t/how-to-access-blockchain-data-from-aws-s3-parquet-format/1630 "2023-08-29T12:27:44Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Divya](https://avatars.discourse-cdn.com/v4/letter/d/13edae/32.png) [@Divya](https://community.bitquery.io/u/Divya)\
**Post date:** [August 29, 2023, 12:27pm UTC](https://community.bitquery.io/t/how-to-access-blockchain-data-from-aws-s3-parquet-format/1630/1 "2023-08-29T12:27:45Z")

</div>

Bitquery earlier introduced our new offering “Blockchain Data on the Cloud” where you can have any blockchain data uploaded to AWS.  
[In the previous article](https://community.bitquery.io/t/how-to-access-blockchain-data-from-aws-s3/1456) we saw how to extract blockchain data from AWS S3 buckets where the data was in protobuf schema.  
This article will focus on extracting blockchain data from AWS S3 buckets in parquet format.

### Step-by-step tutorial :

1. Install and Import the necessary libraries.

```Python
import boto3
import pyarrow as pa
import pyarrow.parquet as pq
import pandas as pd 
import parquet

```

The `boto3` library allows us to interact with Amazon S3, the `pyarrow` library allows us to read and write Parquet files, the `pandas` library allows us to manipulate DataFrames, and the `parquet` library provides additional functions for working with Parquet files.

1. Set up your AWS credentials.

For this step, you need to get your AWS Access Keys.

- Go to your AWS Console [AWS Management Console](https://aws.amazon.com/console/)
- Navigate to your profile → Security credentials → Generate Access Key  
 ![image](https://us1.discourse-cdn.com/flex016/uploads/bitquery/original/1X/5fe6fae67ae6c4e36ec41986739fa46d86a497b6.png)

```Python
s3 = boto3.client('s3', aws_access_key_id='your id', aws_secret_access_key='your key', region_name='us-east-1')

```

Replace `your id` and `your key` with your actual AWS access key ID and secret access key. You can find these values in the AWS Management Console.

1. Specify the S3 bucket and file name.

```Python
bucket_name = 'parquet-bsc'
bsc_key='dex_trades_tx/2020-09-12_410000_88573E9FA803AF8D_10000.parquet'

```

The above mentioned bucket details are from the sample.  
Replace `parquet-bsc` with the name of your S3 bucket and `dex_trades_tx/2020-09-12_410000_88573E9FA803AF8D_10000.parquet` with the name of the Parquet file you want to download.

1. Download the Parquet file to a local path.

```Python
bsc_local_path = 'C:/Your Path/sample.parquet'

file_obj = s3.download_file(bucket_name, bsc_key, bsc_local_path)

```

The `download_file()` method will download the Parquet file from the S3 bucket to the local path you specified.

1. Read the Parquet file into a Pandas DataFrame.

```Python
df = pd.read_parquet(bsc_local_path)

```

The `read_parquet()` method will read the Parquet file into a Pandas DataFrame.

1. Write the DataFrame to a CSV file.

```Python
df.to_csv('parquet_output.csv', index=False)

```

The `to_csv()` method will write the Pandas DataFrame to a CSV file. The `index=False` argument tells the method not to write the DataFrame index to the CSV file.