BUG: fix handling of hashes in file paths #3280
Conversation
geopandas/io/tests/test_file.py
Outdated
| @@ -694,6 +695,17 @@ def test_allow_legacy_gdal_path(engine, nybb_filename): | |||
| assert isinstance(gdf, geopandas.GeoDataFrame) | |||
|
|
|||
|
|
|||
| def test_read_file_with_hash_in_path(engine, nybb_filename, tmp_path): | |||
| if engine == "pyogrio": | |||
| pytest.xfail("upstream bug") | |||
There was a problem hiding this comment.
we can turn this into a conditional skip once geopandas/pyogrio#412 is in.
m-richards
changed the title
This bug is present in released geopandas 0.14.4 |
Ah good catch, it was backported. To your question @jorisvandenbossche I wasn't 100% sure above, but it looks like this was fine in previous versions with fiona, as it checks if a path string conforms to be below before elif re.match("^[a-z0-9\\+]*://", path):
parts = urlparse(path)I suppose there's the question of if there should be an equivalent check in the pyogrio code / the port of the pyogrio code in geopandas used for the fiona 1.10 support path. |
CHANGELOG.md
Outdated
| @@ -1,6 +1,6 @@ | |||
| # Changelog | |||
|
|
|||
| ## Version 1.0.0-alpha1 (Apr 13, 2024) | |||
| ## Version 1.0.0 | |||
There was a problem hiding this comment.
I'm assuming we deliberately chose to mix new entries into the changelog for alpha1 to create a consolidated changelog for the final 1.0 release, so I updated the header accordingly.
Closes #3279.
This is an upstream bug in pyogrio is well. It may also be upstream in fiona, but haven't check that yet.
Test failure is for dev and unrelated.
There's no changelog entry as this modifies #3232 which is also unreleased (well it's in alpha 1. It seems that we've merged changelog entries after the alpha into the alpha 1 section, so I assume we will actually not keep the alpha section, we'll just have a combined changelog for 1.0)