Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Lance scalar index search loads dataset metadata (which should be cached) #2313

Closed
westonpace opened this issue May 8, 2024 · 0 comments · Fixed by #2330
Closed

Lance scalar index search loads dataset metadata (which should be cached) #2313

westonpace opened this issue May 8, 2024 · 0 comments · Fixed by #2330

Comments

@westonpace
Copy link
Contributor

Tracking the I/O of a scalar index search we can see that the search is loading the dataset metadata (it appears to do so twice). That data should already be cached. Loading it can be quite costly and defeat the purpose of doing an indexed search in the first place. This gets even worse when there are many fragments in a dataset because the manifest is quite large.
@westonpace westonpace mentioned this issue May 13, 2024
Merged
westonpace added a commit that referenced this issue May 15, 2024
@westonpace
perf: use the file metadata cache in scalar indices (#2330)
dfb531b 
We were not using the cache before which meant every scalar index query
had to load a bunch of metadata from the index files (some of which was
quite expensive)

Closes #2313
eddyxu pushed a commit that referenced this issue May 16, 2024
@westonpace @eddyxu
perf: use the file metadata cache in scalar indices (#2330)
06dac46 
We were not using the cache before which meant every scalar index query
had to load a bunch of metadata from the index files (some of which was
quite expensive)

Closes #2313
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

perf: use the file metadata cache in scalar indices westonpace/lance
1 participant
@westonpace