Redshift Spectrum: Automatically partition tables

问题:

We currently generate a daily CSV export that we upload to an S3 bucket, into the following structure:

<report-name>
|--reportDate-<date-stamp>
    |-- part0.csv.gz
    |-- part1.csv.gz

We want to be able to run reports partitioned by daily export.

According to this page, you can partition data in Redshift Spectrum by a key which is based on the source S3 folder where your Spectrum table sources its data. However, from the example, it looks like you need an ALTER statement for each partition:

alter table spectrum.sales_part
add partition(saledate='2008-01-01') 
location 's3://bucket/tickit/spectrum/sales_partition/saledate=2008-01/';

alter table spectrum.sales_part
add partition(saledate='2008-02-01') 
location 's3://awssampledbuswest2/tickit/spectrum/sales_partition/saledate=2008-02/';

Is there any way to set the table up so that data is automatically partitioned by the folder it comes from, or do we need a daily job to ALTER the table to add that day's partition?

回答1:

Solution 1:

At max 20000 partitions can be created per table. You can create a one-time script to add the partitions (at max 20k) for all the future s3 partition folders.

For eg.

If folder s3://bucket/ticket/spectrum/sales_partition/saledate=2017-12/ doesn't exist, you can even add partition for that.

alter table spectrum.sales_part
add partition(saledate='2017-12-01') 
location 's3://bucket/tickit/spectrum/sales_partition/saledate=2017-12/';

Solution 2:

https://aws.amazon.com/blogs/big-data/data-lake-ingestion-automatically-partition-hive-external-tables-with-aws/

回答2:

Another precise way to go about it: Create a Lambda job that is triggered on the ObjectCreated notification from the S3 bucket, and run the SQL to add the partition:

alter table tblname ADD IF NOT EXISTS PARTITION (partition clause) localtion s3://mybucket/localtion