Compression of parquet files in s3 bucket

Question

I have a date pipeline, which takes data from rudderstack and puts it into S3 datalake. In a short time period, my 80k events are taking up 2.3GB.

Is there some way I can do compression on those objects? If possible, what I want to acheive is - data coming to S3 is compressed automatically to reduce storage charges.

Riku_Kobayashi · Answer

Hello.

Since S3 alone does not have a compression feature, it would be a good idea to create a mechanism where a file is compressed using Lambda or Glue when it is uploaded to S3 and then uploaded again.  
When using Lambda, it can be called from S3 with event notification.  
https://docs.aws.amazon.com/AmazonS3/latest/userguide/EventNotifications.html

In the case of AWS Glue, this is possible by using EventBridge etc. to execute the job.  
https://repost.aws/questions/QUsQRxOITUR_qj2uAe3_NUoA/trigger-glue-job-from-s3

Compression of parquet files in s3 bucket

Relevant content