# Resampling Zarr-backed Xarray wth VRTs and GDAL support for Zarr

**URL:** <https://discourse.pangeo.io/t/resampling-zarr-backed-xarray-wth-vrts-and-gdal-support-for-zarr/3829>\
**Category:** Uncategorized\
**Created:** [November 9, 2023, 4:55pm UTC](https://discourse.pangeo.io/t/resampling-zarr-backed-xarray-wth-vrts-and-gdal-support-for-zarr/3829 "2023-11-09T16:55:00Z")\
**Posts on this page:** 1\
**Showing post:** 6

<div class="post-metadata">

**Author:** ![Leonard\_Strnad](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.pangeo.io/leonard_strnad/32/1307_2.png) [@Leonard\_Strnad](https://discourse.pangeo.io/u/Leonard_Strnad)\
**Post date:** [November 10, 2023, 6:18pm UTC](https://discourse.pangeo.io/t/resampling-zarr-backed-xarray-wth-vrts-and-gdal-support-for-zarr/3829/6 "2023-11-10T18:18:23Z")

</div>

It actually looks like this works.

```auto
gdalbuildvrt -tr .1 .1 test.vrt "vrt://ZARR:/vsigs/store/group:/variable_1:1?a_ullr=-180,90,180,-90"

```

where I noticed dropping `a_ullr` results in a `Warning 1: gdalbuildvrt does not support positive NS resolution. Skipping`.

Then, I can open with

```auto
rioxarray.open_rasterio("test.vrt", chunks="auto")

```

This might do the trick! I’ll try on a large-ish dataset and check the data and report back.

It would be sick if the same thing worked for multidimensional vrt and we only need to make one “vrt” for the entire xarray dataset, but I dead end with `rioxarray` saying the file format is not supported.

---

_[View the full topic](https://discourse.pangeo.io/t/resampling-zarr-backed-xarray-wth-vrts-and-gdal-support-for-zarr/3829)._