# Duplicate Vectors and HNSW Memory Usage

**URL:** <https://forum.opensearch.org/t/duplicate-vectors-and-hnsw-memory-usage/6811>\
**Category:** k-NN\
**Created:** [August 18, 2021, 7:02pm UTC](https://forum.opensearch.org/t/duplicate-vectors-and-hnsw-memory-usage/6811 "2021-08-18T19:02:47Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![dnock](https://avatars.discourse-cdn.com/v4/letter/d/919ad9/32.png) [@dnock](https://forum.opensearch.org/u/dnock)\
**Post date:** [August 18, 2021, 7:02pm UTC](https://forum.opensearch.org/t/duplicate-vectors-and-hnsw-memory-usage/6811/1 "2021-08-18T19:02:47Z")

</div>

I was wondering what if anything changes when multiple documents reference the same vector (i.e. doc\_1 = { vector : [1, 2]} and doc\_2 = { vector : [1, 2]})? Specifically, does the memory footprint of the HNSW graph still follow the estimate model listed in the [docs](https://opendistro.github.io/for-elasticsearch-docs/docs/knn/performance-tuning/#estimating-memory-usage).

---

<div class="post-metadata">

**Author:** ![jmazane](https://avatars.discourse-cdn.com/v4/letter/j/f4b2a3/32.png) [@jmazane](https://forum.opensearch.org/u/jmazane)\
**Post date:** [August 18, 2021, 9:53pm UTC](https://forum.opensearch.org/t/duplicate-vectors-and-hnsw-memory-usage/6811/2 "2021-08-18T21:53:56Z")

</div>

Hi @dnock, no nothing changes. Docs are independent so duplicates will exist.

Jack