correct sequence for running maintenance steps on an iceberg table

[salimpadela]

Query engine

Spark, AWS Glue

Question

What is the correct sequence of maintenance steps to run on an Iceberg table? Our tables are write-once-read-many so I am not sure if I need to run rewrite_position_delete_files or not.

I am not very experienced in Iceberg so I wanted to see if I am running them in the correct order or am I missing something here:

Right now my sequence looks like this:

1. rewrite_data_files
2. rewrite_manifests
3. expire_snapshots

I read it on a blog that I don't have to run rewrite_manifests before or after running rewrite_data_files as rewrite_data_files rewrites manifests but I am finding the results contradictory based on what I am seeing.

With the sequence I specified, this is how the results look like for rewrite_data_files and rewrite_manifests calls.

Result of rewrite_data_files

rewritten_data_files_count	added_data_files_count	rewritten_bytes_count	failed_data_files_count
13507	2371	4307611669	0

---

Result of rewrite_manifests

rewritten_manifests_count	added_manifests_count
29	25

P.S. My intention is to have only one snapshot at the end of the process so I am providing older_than in expire_snapshots.