# Loading dataset from cache .arrow file

**URL:** <https://discuss.huggingface.co/t/loading-dataset-from-cache-arrow-file/34932>\
**Category:** 🤗Datasets\
**Created:** [March 28, 2023, 5:56pm UTC](https://discuss.huggingface.co/t/loading-dataset-from-cache-arrow-file/34932 "2023-03-28T17:56:29Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![p4vv37](https://avatars.discourse-cdn.com/v4/letter/p/8e8cbc/32.png) [@p4vv37](https://discuss.huggingface.co/u/p4vv37)\
**Post date:** [March 28, 2023, 5:56pm UTC](https://discuss.huggingface.co/t/loading-dataset-from-cache-arrow-file/34932/1 "2023-03-28T17:56:29Z")

</div>

Hi

In my script I did some computationally costly processing of a large dataset and performed .save\_to\_disk on it. Unfortunately, it failed on that step, the dataset was not saved.

I still have the gigantic cache-4d6878f4ef3ef2f5.arrow file on the drive, am I able to load this cached version of dataset?

Regards

---

<div class="post-metadata">

**Author:** ![p4vv37](https://avatars.discourse-cdn.com/v4/letter/p/8e8cbc/32.png) [@p4vv37](https://discuss.huggingface.co/u/p4vv37)\
**Post date:** [March 28, 2023, 6:00pm UTC](https://discuss.huggingface.co/t/loading-dataset-from-cache-arrow-file/34932/2 "2023-03-28T18:00:02Z")

</div>

I found the answer:

```auto
Dataset.from_file

```

😮‍💨