Data Explorer: Compute column profiles in smaller batches for large data frames, avoid timeouts #5148
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
… than in a monolithic batch
wesm
changed the title
|
Latest change with batches of 4 (which seems to strike a good balance of keeping the backend busy without doing too much work at once): Screencast.from.2024-10-24.11-01-33.mp4The double compute bug is visible here (and not caused by these changes) so I'll take a closer look at that |
wesm
added a commit
that referenced
this pull request
Oct 24, 2024
#5156) Attempts to address #5150. In the backend-state-updated event handler, the profiles were being refreshed only before `fetchData` was called with an argument to invalidate all caches including the profiles, so the profiles have to immediately be recomputed. This resolves the double computation that I observed in #5148.
|
Merging this, we can keep improving the small details in follow up work |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Addresses #4629 and #2851. This fixes timeouts that were occurring, making it seem like the data explorer is broken on large datasets, while improving the performance / responsiveness by computing updates incrementally rather than in a monolithic batch which may take 10s of seconds to return.
I found that if we spam the backend with all of the profile requests at once, the
get_data_valuesrequest doesn't get served because of all the thread contention, so this just has one profile request active a time.You can see what a 33M row dataset looks like on the initial open and then various filtering workflows
Screencast.from.2024-10-23.17-09-33.mp4
This does make the lack of a loading indicator somewhat more pronounced, but that can be addressed separately.
If you watch the end of the video, you can see that there is a bug where after the profiles are all computed, the cache is cleared and they are computed again. This bug appears to be present even without this change so I will open another issue (see #5150).