From 7d491e1f88dd4710eb1056a9e4b38e2403b9b63b Mon Sep 17 00:00:00 2001 From: Dean Wyatte <2512762+dwyatte@users.noreply.github.com> Date: Fri, 10 Mar 2023 14:18:01 -0700 Subject: [PATCH] Update src/datasets/utils/file_utils.py Co-authored-by: Quentin Lhoest <42851186+lhoestq@users.noreply.github.com> --- src/datasets/utils/file_utils.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/datasets/utils/file_utils.py b/src/datasets/utils/file_utils.py index 4b9ed454868b..00d6c00b2188 100644 --- a/src/datasets/utils/file_utils.py +++ b/src/datasets/utils/file_utils.py @@ -501,7 +501,8 @@ def get_from_cache( elif scheme not in ("http", "https"): response = fsspec_head(url) # use the hash of the response as a pseudo ETag to detect changes - etag = json.dumps(response, sort_keys=True) if use_etag else None + # s3fs uses "ETag", gcsfs uses "etag" + etag = (response.get("ETag", None) or response.get("etag", None)) if use_etag else None connected = True try: response = http_head(