Enhance Layer Drawing Speed in ArcGIS Pro with Snowflake Data Source

[SBark_CBRE]

Hello Esri Community - I am reaching out to seek your advice on optimising layer drawing speed in ArcGIS Pro when connected to a Snowflake data source.

I am currently working on a project where we are using Snowflake as our primary data storage, and we need to visualise and interact with large datasets in ArcGIS Pro. However, we have been facing massive performance issues, as the layers are taking a significant amount of time to render and update (15+ seconds). The data is small and not complex (10,000 points, ABP data) and when we query the data I can see Snowflake is responding in under 1 second, so it's Esri side that is slowing it down (roughly 15-20 seconds to draw). Same experience when the layer is published to Enterprise, still very slow.

Here is a brief overview of our setup and what we have tried so far:

1. ArcGIS Pro Version: 3.0
2. Snowflake Connector: ODBC Driver (64 bit)
3. Data Structure: Address Base Premium (OS) point data
4. System Specifications: Windows 10 Enterprise with 16gb RAM

We have already attempted the following approaches to improve the drawing speed:

1. Applied spatial indexing on the relevant Snowflake tables.
2. Reduced the number of features being drawn by implementing definition queries.
3. Optimized map scale settings and display settings.
4. Ensured that the system meets the minimum hardware requirements for ArcGIS Pro.

Unfortunately, despite these efforts, we are still experiencing slow drawing speeds. If anybody has any known limitations or workarounds for connecting Snowflake data sources with ArcGIS Pro, it'd be much appreciated.

Thanks in advance,

Sam 

[VaibhavS]

ArcGIS Pro > Settings > Options > Map and Scene options > under Add Layer section > uncheck ' Make newly added layers visible by default' 

Go to Geoprocessing options > uncheck ' Allow geoprocessing tools to overwrite existing datasets.

In the same group, under Logging > uncheck 'Write geoprocessing operations to Geoprocessing history and 'Write geoprocessing operations to dataset metadata.

Save the settings and try to load your data. 

Until and unless you don't want the geoprocessing history and metadata saved, we can uncheck these. 

[SBark_CBRE]

Thanks for the reply but unfortunately no luck 😞 

[JonathanPollack]

did you ever get this resolved, we are seeing similar HORRIBLE performance

[RobertAkroyd1]

I've been doing a bit of testing on this and have pretty good performance using ArcGIS Pro (3.2.2).  I loaded GB Address Point data - 30.4 million points into Snowflake.  I have an X-Large warehouse and importantly my Snowflake is hosted in the UK - EU (London) as that is where I am too using ArcGIS Pro.

I enabled Binning on the dataset using ArcGIS Pro, and at small scales the bins work well to aggregate the data:

Then when zoomed in beyond the binning threshold it draws the individual addresses pretty quickly:

I'm doing more testing and will post back here at some point.