Comments (6)
Hi,
I'm looking at the Iceberg tables support plan:
As iceberg doesn't support CTA, the implementation do the following:
create tmp table as parquet
drop the old table, if exist
create iceberg table based on tmp table definition, only metadata
insert into from tmp
drop tmp table
I think that the tmp table creation is going to kill performance (I have 10s of billions of rows).
How about this approach:
create the tmp table with the model's SELECT with "LIMIT 0" appended
create iceberg table based on tmp table definition, only metadata
insert into the iceberg table with the model's SQL (don't append the LIMIT this time)
drop the tmp table
This would save one copying of all rows (hours of processing in my case).
Regards,
ZD
from dbt-athena.
@zsvoboda I implemented what you suggested, I tested with small datasets, seems working fine.
from dbt-athena.
Thank you!!! I’ll test it as soon as possible. What branch should I use for testing it?
from dbt-athena.
@zsvoboda we released the changes to 1.3.1, try that out.
from dbt-athena.
Hi Nicola, I'm testing the 1.3.2, and it works great in the table materialization mode. My elt works much faster now. Thank you! However, it works the old way in the incremental materialization mode. Any chance you can do the same trick in the incremental mode?
from dbt-athena.
@zsvoboda Feel free to address another issue to handle that request. I believe that is possible:
- we first use parquet with limit 0
- do the schema evolution
- use the merge from the original SQL
from dbt-athena.
Related Issues (20)
- [Bug] Compilation Error when create view with lf-tags
- [Bug] pre_hook executed two times HOT 2
- [Feature] Table name prefix via profile configuration HOT 3
- [Bug] adapter response return incorrect `data_scanned_in_bytes` when incremental model is running HOT 3
- Docs - Add bucketing documentation for iceberg
- Upgrade to support dbt-core v1.8.0 HOT 1
- [Bug] `truncate()` partition transformation does not work when it includes more than 100 partitions HOT 1
- Bug Hitting `ThrottlingException` on `GetWorkGroup` with threads turned up HOT 5
- [Bug] Iceberg table materialization shouldn't s3_data_naming=table
- [Bug] Adapter error when FIPS mode is enabled HOT 4
- [Bug] Resolution failure for `create_table_as` macro when upgrading to 1.7.2 HOT 1
- upgrade to support dbt-core v1.8.0 HOT 6
- [Feature] Control glue database/schema for tmp tables generated by incremental models
- [Bug] force_batch deletes data from model_tmp_not_partitioned before coping to the final table HOT 2
- [Feature] Rename unique_key to unique_columns or merge_on_columns HOT 3
- [Feature] Support configurable management of Table Optimisers for Iceberg tables HOT 2
- [Bug] Error when Python Model Goes To Write To Database HOT 13
- [Feature] Custom strategy for incremental models when table type is iceberg
- [Bug] dbt source freshness expected a timestamp but received a string HOT 1
- [Feature] Athena dbt-external-tables impl as independent package HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from dbt-athena.