You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Aug 29, 2025. It is now read-only.
Repository navigation
This repository was archived by the owner on Aug 29, 2025. It is now read-only.
Scatter with animation_frame is plotted out of order #23
If you're using Python 3 then this is a bug :) Can you post a minimal reproducible example in source here please?
the category_orders argument is a dict where the keys are column names and the values are lists containing the values you want to see in the order you want to see them. So in this case you'd have category_orders={"frame": [0,1,2,3,...]}
Ah, sorry, I should be clearer: if you're using Python below 3.7 (i.e. not just 2.*) then order is not guaranteed. As of Python 3.7 iterating over dict keys is guaranteed to be in insertion order
where 0 is the column name (use quotes if your column has an actual name) and range can be replaced with the list of all the items of that particular column but in the order you want.
@nicolaskruchten I cannot share the file right now, but it's just a CSV file in the following format"
index,value1x,value1y,value2x,value2y....
I'm really struggling to replicate this bug on this end with Python 3.7. In fact, I'm struggling to get the frames to render out of order at all no matter what I do to the data frame :)
OK, I've just pushed what I believe is a fix for this in version 0.2.1, such that animation frames by default now should always render in the order the values are encountered in the input data frame. Please let me know if you're still encountering issues with this latest version!
Wow, that's really bizarre. Can you share your data as well so that I can replicate this locally? What happens if you don't set category_orders at all with v0.2.1?
OK, thanks for sharing this data! I can reproduce your problem on this end and I'll try to push out another fix shortly. I really appreciate the patience and follow-through: this makes the library better for everyone 😄 🙏
Activity
nicolaskruchten commentedon Mar 22, 2019
If you're using Python 2, then order is not guaranteed unfortunately and you'll have to use the
category_ordersargument to specify the order for now.dvduvix commentedon Mar 22, 2019
@nicolaskruchten I am using Python 3. But for the sake of trying: how do I use category_orders?
nicolaskruchten commentedon Mar 22, 2019
If you're using Python 3 then this is a bug :) Can you post a minimal reproducible example in source here please?
the
category_ordersargument is a dict where the keys are column names and the values are lists containing the values you want to see in the order you want to see them. So in this case you'd havecategory_orders={"frame": [0,1,2,3,...]}nicolaskruchten commentedon Mar 22, 2019
Ah, sorry, I should be clearer: if you're using Python below 3.7 (i.e. not just 2.*) then order is not guaranteed. As of Python 3.7 iterating over dict keys is guaranteed to be in insertion order
dvduvix commentedon Mar 22, 2019
@nicolaskruchten Then that must be it. I have Python 3.5.2
kmader commentedon Mar 22, 2019
Thanks for that insight about Python 3.5.2 vs 3.7 I hadn't thought of that. Here is the minimal code/environment to run it https://gist.github.com/kmader/7a182744194a5330008b6e40d60aea6b and the link to just run it on mybinder: https://mybinder.org/v2/gist/kmader/7a182744194a5330008b6e40d60aea6b/master?filepath=example.ipynb
As Kaggle Kernels use Python >3.5 it works fine on their site: https://www.kaggle.com/kmader/plotly-express-bug
dvduvix commentedon May 2, 2019
@nicolaskruchten So I finally decided to test the theory and installed Python 3.7.3. The out of order problem remains.
The only thing that worked to fix this was using the category_orders.
nicolaskruchten commentedon May 8, 2019
@dvduvix this is very puzzling... are you able to share a CSV of your data or some other way I could reproduce this locally?
nicolaskruchten commentedon May 8, 2019
(also, perhaps try with v0.1.8 which just came out with a subtle change that might just shake this problem loose)
MengyinZhong commentedon May 9, 2019
@dvduvix
Hi, I met the same issue and I attempted to fix it, but it doesn't seem to work, so could you please post your code? Thx a lot~😃
dvduvix commentedon May 9, 2019
@MengyinZhong
All I did is added a category_orders to scatter.
px.scatter(data, animation_frame='frame', animation_group='col_name', category_orders={0: range(1000)})
where 0 is the column name (use quotes if your column has an actual name) and range can be replaced with the list of all the items of that particular column but in the order you want.
@nicolaskruchten I cannot share the file right now, but it's just a CSV file in the following format"
index,value1x,value1y,value2x,value2y....
I will try the latest version as soon as I can
nicolaskruchten commentedon May 9, 2019
I'm really struggling to replicate this bug on this end with Python 3.7. In fact, I'm struggling to get the frames to render out of order at all no matter what I do to the data frame :)
MengyinZhong commentedon May 10, 2019
@dvduvix Thank you for your reply, and I tried the solution you provided, but the timeline is still out of order, shown as below
nicolaskruchten commentedon May 10, 2019
@MengyinZhong what version of Python and plotly_express are you using?
fix groupby ordering when multi-grouping, should address #23
nicolaskruchten commentedon May 13, 2019
OK, I've just pushed what I believe is a fix for this in version 0.2.1, such that animation frames by default now should always render in the order the values are encountered in the input data frame. Please let me know if you're still encountering issues with this latest version!
MengyinZhong commentedon May 13, 2019
@nicolaskruchten
Python 3.7.0
plotly_express 0.2.1
Sorry for my tardy reply, but the order still cannot show as expected:
nicolaskruchten commentedon May 13, 2019
Wow, that's really bizarre. Can you share your data as well so that I can replicate this locally? What happens if you don't set
category_ordersat all with v0.2.1?MengyinZhong commentedon May 13, 2019
@nicolaskruchten I got the same result without setting category_orders. Here is my data https://github.com/MengyinZhong/test/blob/master/test.csv
nicolaskruchten commentedon May 13, 2019
OK, thanks for sharing this data! I can reproduce your problem on this end and I'll try to push out another fix shortly. I really appreciate the patience and follow-through: this makes the library better for everyone 😄 🙏
final frame order enforcement, #23
nicolaskruchten commentedon May 13, 2019
OK, I think I've got it for real this time! Please try v0.2.2, with or without
category_orders, on any version of Python :)nicolaskruchten commentedon Jun 3, 2019
Closing for now, please reopen if this happens again :)
ariewjy commentedon Jul 18, 2023
Hi, @nicolaskruchten ,
I have the same problem, the year animation frame is out of order, and I am having the same issue in my laptop and also in Kaggle notebook.
Here is the link to the notebook to replicate the issue.
https://www.kaggle.com/ariewijaya/plotly-animation-bug
I am using latest plotly, same as in Kaggle Notebook.
Thanks!