Where is the full dataset?

[moore269]

I see there are 178 examples in this file
https://github.com/xlang-ai/Spider2/blob/main/spider2/examples/spider2.jsonl

However, the paper says there are 600 examples? Where are the rest? Also, is it possible to have the correctly labeled sql as another field in the jsonl as well?

Lastly, is there a place where I can easily download all the referenced tables?

[lfy79001]

Hi,

We are working on 600 examples, and we have currently only released part of the data. It is expected to take another week.

Some of the tables are on the cloud, please refer to Bigquery Guideline, so you don’t need to download them. There is another portion of the tables that need to be downloaded, which you can access via this link

[sethsiddharth]

Hello @lfy79001!
It's been about 3 weeks since the last update on the full dataset release. Any news on the progress?

Thank you for your work on this project and for keeping the community informed.

[lfy79001]

Thank you for your interest in Spider 2.0. We have been busy with paper writing and data validation. In about 10 days, we will release the paper and all the data.

[sethsiddharth]

Thank you for the update. Looking forward to the release!

[fi5421]

Hello,
Has the complete dataset for nlp questions released? I am interested in the spider2.0-snow dataset which should have 547 questions however the spider2-snow.jsonl only has 260 questions.

[lfy79001]

Thank you for your interest in our work.
Spider 2.0 , Spider 2.0-lite and Spider 2.0-snow, have now been released with 50% dev split dataset. However, we currently do not plan to release the full dataset of questions to ensure the fairness of the competition.

At this stage, we encourage you to follow the guidelines provided in this document: Submission Guidelines. We will assist with the evaluation of your method based on these instructions.

We may consider releasing all questions in a few months. If there are any changes to the dataset release plan, I will inform you promptly.

Best regards

[fi5421]

Thanks for the update!