Comments (6)
It seems like you're probably running into the zetasql max_intermediate_byte_size. It's set in-code here, and at the moment there's no way to configure it except through code changes. We would need to submit a feature request to zetasql to make it configurable, and then update our query execution tool to expose the option.
As a short-term fix, you could try editing the limit in code. You might also be able to revise your query to avoid the limit - what query are you trying to run?
from differential-privacy.
Thank you for your reply.
The query is quite simple, just a sum() of one column y a dataset with more than 150k-200k rows. Not much to simplfy, apart from performing multiple sums o paritions of the dataset.
Will try to modify the code until its possible to make it configurable.
Is there an official way of requesting the feature to zetasql, apart from a github issue?
from differential-privacy.
I filed an internal feature request with them on your behalf, so no need for you to do anything there :)
I'll update this issue when we have progress.
from differential-privacy.
Thank you very much!
from differential-privacy.
Thanks for your patience! We have now updated the ZetaSQL dependencies and you could now try to increase memory by setting the following flags:
--evaluator_max_value_byte_size
--evaluator_max_intermediate_byte_size
Defaults are:
int64_t max_value_byte_size = 1024 * 1024;
int64_t max_intermediate_byte_size = 128 * 1024 * 1024;
Let me know if this works for your use case.
from differential-privacy.
Closing this, as we have a workaround with the above flags. Documentation changes are in the pipeline to be pushed soon.
from differential-privacy.
Related Issues (20)
- Use DP library in make project HOT 4
- README code fencing is off by one. HOT 2
- Build fails on Mac HOT 6
- Is there a plan to address the issues pointed out in Casacuberta et al? HOT 2
- > Is there a plan to address this:loating-point implementations are subject to the vulnerabilities described in [Casacuberta et al. "Widespread Underestimation of Sensitivity in Differentially Private Libraries and How to Fix it"](https://arxiv.org/abs/2207.10635) (specifically the rounding, repeated rounding, and re-ordering attacks). HOT 1
- Updated Maven artifacts? HOT 6
- Google Colab ? HOT 3
- SQL Queries HOT 4
- Why the privacy-on-beam is not implemented by java? HOT 1
- Implementing a custom noisy threshold HOT 3
- Making the Introduction of the Readme Section Easier to Read and Understand HOT 3
- Do ZetaSQL examples supports JOIN queries? HOT 4
- Invalid uses of `list` type HOT 1
- Invalid use of `list` type should be `typing.List` HOT 5
- Relax dependency for scipy~=1.7.1 in dp_accounting HOT 2
- run `bazelisk build execute_query` failed HOT 5
- bazel build excute_query with "linkopts = ["-static"]", and the output binary does not work as expected HOT 22
- Maven artifact with LongBoundedSum HOT 3
- postgres extension docker build failed HOT 4
- install postgres extension failed with `./postgres/install_extension.sh` HOT 2
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 differential-privacy.