# Efficiently gather a complete GBIF dataset?

**URL:** <https://discourse.gbif.org/t/efficiently-gather-a-complete-gbif-dataset/6423>\
**Category:** API developers\
**Created:** [August 12, 2026, 8:02am UTC](https://discourse.gbif.org/t/efficiently-gather-a-complete-gbif-dataset/6423 "2026-08-12T08:02:45Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![jervenbolleman](https://avatars.discourse-cdn.com/v4/letter/j/ac8455/32.png) [@jervenbolleman](https://discourse.gbif.org/u/jervenbolleman)\
**Post date:** [August 12, 2026, 8:02am UTC](https://discourse.gbif.org/t/efficiently-gather-a-complete-gbif-dataset/6423/1 "2026-08-12T08:02:45Z")

</div>

Hi All,

We created a fast GBIF parquet to RDF converter[0] and some people like the capabilities that gives. We have one user asking for some information not available in the GBIF parquet data [1].

What is the recommended way to download all GBIF data? should we retrieve each DWC archive one by one via API or is there a cheaper way to do this?

Regards,

Jerven

PS. The LLMs suggest DWC-A download, but that would be expensive if it is a download of all of GBIF.

[0] [GitHub - Micelio/gbif\_parquet · GitHub](https://github.com/Micelio/gbif_parquet/)

[1] [GBIF data that would be cool to have that is not in the SIMPLE parquet format · Issue #13 · Micelio/gbif\_parquet · GitHub](https://github.com/Micelio/gbif_parquet/issues/13)

---

<div class="post-metadata">

**Author:** ![estebanmhGBIF](https://avatars.discourse-cdn.com/v4/letter/e/f07891/32.png) [@estebanmhGBIF](https://discourse.gbif.org/u/estebanmhGBIF)\
**Post date:** [August 12, 2026, 1:19pm UTC](https://discourse.gbif.org/t/efficiently-gather-a-complete-gbif-dataset/6423/2 "2026-08-12T13:19:22Z")

</div>

Hi @jervenbolleman

If you want the raw data, you can use the [Monthly snapshot datasets in our Occurrence snapshots](https://www.gbif.org/occurrence-snapshots#:~:text=Monthly%20snapshot%20datasets) in the Darwin Core Archive format. There are several options with slightly different queries (with coordinates, occurrenceStatus: present, etc), but if you want everything you can look for the ‘No filters used’ Query: [https://doi.org/10.15468/dl.8yrbe7](https://doi.org/10.15468/dl.8yrbe7) that will also have all the extensions and you can use that for your converter.

It is a big file of 2TB, but that is just because of the size of the data, even the simple [download](https://doi.org/10.15468/dl.eh6a2p) is 465 GB.

Let us know if this solves your question or do you want to know something else

---

<div class="post-metadata">

**Author:** ![jervenbolleman](https://avatars.discourse-cdn.com/v4/letter/j/ac8455/32.png) [@jervenbolleman](https://discourse.gbif.org/u/jervenbolleman)\
**Post date:** [August 12, 2026, 2:05pm UTC](https://discourse.gbif.org/t/efficiently-gather-a-complete-gbif-dataset/6423/3 "2026-08-12T14:05:34Z")

</div>

Thanks the complete 2TB no filter option is what we will need. Good to know we wont be overloading some API endpoint!