# Can we make virtual dataset with non atomic datatype

**URL:** <https://forum.hdfgroup.org/t/can-we-make-virtual-dataset-with-non-atomic-datatype/12192>\
**Category:** h5py\
**Created:** [April 12, 2024, 1:42pm UTC](https://forum.hdfgroup.org/t/can-we-make-virtual-dataset-with-non-atomic-datatype/12192 "2024-04-12T13:42:20Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![r.trebu](https://forum.hdfgroup.org/letter_avatar_proxy/v4/letter/r/e36b37/32.png) [@r.trebu](https://forum.hdfgroup.org/u/r.trebu)\
**Post date:** [April 12, 2024, 1:42pm UTC](https://forum.hdfgroup.org/t/can-we-make-virtual-dataset-with-non-atomic-datatype/12192/1 "2024-04-12T13:42:20Z")

</div>

Since we can’t make compound dataset with bytes type i was wondering if we can create 2 dataset 1 compound 1 with bytes types and “merge” them using virtual dataset.

---

<div class="post-metadata">

**Author:** ![ajelenak](https://forum.hdfgroup.org/letter_avatar_proxy/v4/letter/a/3d9bf3/32.png) [@ajelenak](https://forum.hdfgroup.org/u/ajelenak)\
**Post date:** [April 12, 2024, 6:27pm UTC](https://forum.hdfgroup.org/t/can-we-make-virtual-dataset-with-non-atomic-datatype/12192/2 "2024-04-12T18:27:51Z")

</div>

What do you mean by bytes type? Just a block of bytes?

I don’t think the current Virtual Dataset feature supports virtualization of individual dataset datatypes.

-Aleksandar