Category Archives: CBO

Object Statistics in Your Data Pipelines (Part 2)

For me, building efficient data pipelines means thinking about performance from day one. And that means a conscious and thoughtful approach to optimizer statistics. It is not just this Autotask, which always runs in the database in the evenings or at weekends and is the responsibility of the DBAs. Every data engineer knows the data to be processed better, knows when and how it is loaded and processed. They should also be better able to recognize when and whether the statistics are critical at certain process steps and therefore when and how the statistics should be collected in the best possible way. That’s their job too! This is my point of view, which I would like to explain and emphasize with the posts in this series

Continue reading

Object Statistics in Your Data Pipelines

As my electrical engineering lecturer at university used to joke, there are only two faults in electricity: 1) No contact where it is needed. 2) There is contact where it is not needed. You can often think the same way when looking for the causes of a poor execution plan: no stats were gathered when the optimizer needed them for a good execution plan or someone gathered stats when it was inappropriate.

Continue reading

Issue with the Hint ENABLE_PARALLEL_DML

Performing an ETL with large data sets, it is often a good idea to run DML in parallel. But, in contrast to parallel query or DDL, parallel DML has to be explicitly enabled. You had to issue ALTER SESSION ENABLE PARALLEL DML in the past. Starting with 12c you can enable parallel DML specifically for each query using the hint ENABLE_PARALLEL_DML. For a few years now, I’ve been using the hint now and then and was quite happy. An observation I made a few days ago can lead to a rethinking. What I could observe is that for the SQL with embedded hint a new child cursor was created each time. Let’s test it!

Continue reading

Online Statistics Gathering for ETL – Part 3

Preparing my session on Online Statistics Gathering for ETL for the DOAG conference, I noticed some points that I didn’t covered in the previous two blog posts. The first point is showing the problem that may arise if indexes are involved. The second one is about partition exchange load and it completes the topic of partitioned tables started in part 2. No blog posting on Oracle products is nowadays complete without mentioning the cloud. The third point is about Autonomous Data Warehouse Cloud Service and Online Statistics Gathering improvements. Continue reading