Comments (2)
(1) sounds good to me, actually, so it's fine to not do this here (and I can just create a PR for GroupBy.agg
or something similar).
Edit: basically, let's make use of pydata/xarray#7206
from flox.
It would be a decent bit of complexity to add, and I'm not inclined to add it.
There would be two advantages:
- The data are only factorized once, and the integer codes are reused.
- We could drastically reduce the number of tasks in the dask graph at the cost of more complicated code. Number of tasks is reduced because we can maker a single task calculate all the necessary intermediates for all reductions.
I'm not sure (1) is worth it, at least for xarray, because after pydata/xarray#7206, we will get this for free by just calling each individual method on a saved GroupBy object (for xarray).
I'm not sure (2) is worth it for a couple of cases:
- It will also mean that to calculate
max
only you will calculate every other reduction and then discard it. - If you're writing the output to zarr for example, you lose parallelism again.
- It could be an advantage to only compute
count
once and reuse it forcount
,mean
but not sure its worth it. We could get this advantage by instead breaking up the current algo to. computecount
andsum
separately formean
. Then the dask optimizer will handle the sharedcount
computation for us.
from flox.
Related Issues (20)
- Reporting a vulnerability HOT 1
- always factorize early
- optimize groupby for resample
- Add engine="numbagg"
- AttributeError: 'DataArrayResample' object has no attribute '_unique_coord'
- Test failure on i386: ValueError: bins must be monotonically increasing or decreasing HOT 4
- More Groupers / user stories / strategies HOT 2
- Support xarray grouper objects in xarray interface
- How to create two groups from two lists of dimension labels, and apply "sum" to each group HOT 2
- Using Xarray and Flox for custom non-aggregation functions HOT 3
- use engine flox if array is ordered? HOT 2
- "most common" Aggregator with Dask HOT 7
- Address repo-review comments
- more cohorts optimization when chunksize == 1
- add cftime benchmarks
- Optimize `split_every` HOT 4
- Error when data variables have different dimensions HOT 2
- Flox seems much slower in some cases? HOT 2
- possible support for sparse arrays HOT 2
- Examples in docs can be hard to read in dark mode HOT 1
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 flox.