# Document tags vs dataset tags

**URL:** <https://forum.inductiveautomation.com/t/document-tags-vs-dataset-tags/116361>\
**Category:** Ignition\
**Created:** [July 10, 2026, 8:20pm UTC](https://forum.inductiveautomation.com/t/document-tags-vs-dataset-tags/116361 "2026-07-10T20:20:53Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![ignition\_user243](https://avatars.discourse-cdn.com/v4/letter/i/59ef9b/32.png) [@ignition\_user243](https://forum.inductiveautomation.com/u/ignition_user243)\
**Post date:** [July 10, 2026, 8:20pm UTC](https://forum.inductiveautomation.com/t/document-tags-vs-dataset-tags/116361/1 "2026-07-10T20:20:54Z")

</div>

Hello,

I am wondering if anyone has any insight regarding the performance of document tags vs dataset tags. We have a scenario where we need to store results from a query in a tag. However, the structure of the result set from the query may differ across gateways. Looking to use a document tag for this but from my research it doesn't seem like document tags are very popular. What are the pros and cons of both?

---

<div class="post-metadata">

**Author:** ![pturmel](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/pturmel/32/3295_2.png) [@pturmel](https://forum.inductiveautomation.com/u/pturmel)\
**Post date:** [July 10, 2026, 8:35pm UTC](https://forum.inductiveautomation.com/t/document-tags-vs-dataset-tags/116361/2 "2026-07-10T20:35:46Z")

</div>

Datasets are fundamentally two-dimensional, with fixed column types. This efficiently matches the standard SQL interface. If you are going to hold the result of a DB query, a Dataset is usually the right answer.

Documents are fundamentally unstructured, requiring a bit more metadata to hold the same content as a dataset, but are not limited to a flat, two-dimensional "array". Documents can hold arrays or dictionaries of arrays or dictionaries (...of arrays or dictionaries...) to any desired depth, or variable depth.

---

<div class="post-metadata">

**Author:** ![lane.duncan](https://sea1.discourse-cdn.com/inductiveautomation/user_avatar/forum.inductiveautomation.com/lane.duncan/32/83700_2.png) [@lane.duncan](https://forum.inductiveautomation.com/u/lane.duncan)\
**Post date:** [July 11, 2026, 5:46am UTC](https://forum.inductiveautomation.com/t/document-tags-vs-dataset-tags/116361/3 "2026-07-11T05:46:20Z")

</div>

> [@pturmel](#):
>
> with fixed column types

Emphasis on this. Having a consistent and strong data type up front helps prevent unforeseen script issues due to incorrect Jython duck typing.