add benchmarking scripts for parquet loader - #192
Merged
sam-herman merged 3 commits intoOct 31, 2025
Conversation
Signed-off-by: Akash Shankaran <akash.shankaran@ibm.com>
akash-shankaran
requested review from
jimdickinson,
sam-herman,
sandoichi and
tjake
as code owners
October 27, 2025 21:50
Signed-off-by: Akash Shankaran <akash.shankaran@ibm.com>
sam-herman
approved these changes
Oct 28, 2025
Collaborator
|
@akash-shankaran thank you for the contribution, can you add a testing section in the description with sample of input/output? |
akash-shankaran
force-pushed
the
addJakeScripts
branch
from
October 30, 2025 23:48
91ff2fa to
28892e1
Compare
Signed-off-by: akash shankaran <akash.shankaran1@gmail.com>
Collaborator
Author
@sam-herman done. I also fixed a bunch of bugs in the script when testing with a new parquet file. (such as parsing columns that didn't exist in original file from Jake :) ). |
opensearch-trigger-bot Bot
pushed a commit
that referenced
this pull request
Oct 31, 2025
* add scripts for parquet loader --------- Signed-off-by: Akash Shankaran <akash.shankaran@ibm.com> Signed-off-by: akash shankaran <akash.shankaran1@gmail.com> (cherry picked from commit 4b6fe50)
sam-herman
pushed a commit
that referenced
this pull request
Oct 31, 2025
* add scripts for parquet loader --------- (cherry picked from commit 4b6fe50) Signed-off-by: Akash Shankaran <akash.shankaran@ibm.com> Signed-off-by: akash shankaran <akash.shankaran1@gmail.com> Co-authored-by: Akash Shankaran <akash.shankaran1@gmail.com>
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
This change adds the scripts used to ingest vector data as parquet files to opensearch, when using JVector engine.
Testing
akashshankaran@parquet-loader>>python3 osbench.py --path 0000.parquet --procs 1
Starting 1 processes over 1 files; batch=8000 bulk_chunk=1000 max_chunk_bytes=10485760
[worker] Missing required column 'embeddings' in batch — available columns: []
[worker] Missing required column 'embeddings' in batch — available columns: [] | 0/1 [00:00<?, ?it/s]
Workers: 100%|███████████████████████████| 1/1 [00:00<00:00, 6.89it/s]
Indexing finished.██████████████████████████| 1/1 [00:00<00:00, 6.90it/s]
Files processed: 1
Approx rows (metadata): 15857
Processes: 1
Successfully indexed: 0
Failed: 0
• retryable (suppressed):0
• critical (shown): 0
Total time: 0.15s
Check List
--signoff.By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.