# Write compressed Tables/Matrices

**URL:** <https://discuss.hail.is/t/write-compressed-tables-matrices/3759>\
**Category:** Hail Query & hailctl\
**Created:** [March 6, 2024, 9:39am UTC](https://discuss.hail.is/t/write-compressed-tables-matrices/3759 "2024-03-06T09:39:53Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![cliveno1](https://yyz2.discourse-cdn.com/flex036/user_avatar/discuss.hail.is/cliveno1/32/1115_2.png) [@cliveno1](https://discuss.hail.is/u/cliveno1)\
**Post date:** [March 6, 2024, 9:39am UTC](https://discuss.hail.is/t/write-compressed-tables-matrices/3759/1 "2024-03-06T09:39:53Z")

</div>

I have large genomic (Terabytes of VCFs) files that I wish to write to a database as Hail. When I write with `hl.write` the disk space of the Hail objects is actually around 1.4x bigger than the original VCF files. Is there a more efficient way to write these tables?

---

<div class="post-metadata">

**Author:** ![patrick-schultz](https://yyz2.discourse-cdn.com/flex036/user_avatar/discuss.hail.is/patrick-schultz/32/265_2.png) [@patrick-schultz](https://discuss.hail.is/u/patrick-schultz)\
**Post date:** [March 25, 2024, 4:08pm UTC](https://discuss.hail.is/t/write-compressed-tables-matrices/3759/2 "2024-03-25T16:08:24Z")

</div>

Hi @cliveno1,

That is definitely surprising. Can you share any more details? If you import one of the VCFs as `mt`, what is the output of `mt.describe()`?