I am looking for some opinions on the best way forwards with an interactive perspective page I am building, with regard to handling large datasets returned from a post call to an API.
Site-wide users will enter a production order number and the post call returns all the production order information in JSON format. The data will be displayed in tables, showing operation status etc. Pretty standard so far.
The returned JSON from the ERP system can have thousands of key:value pairs, with various nesting levels.
I have previously built a Vision client that does similar with the same API and using recursive functions I can push the data to dataset tags for my further use. (Removing invalid values, making headers human readable, etc using dataset query expression tags.) All good.
This is where I now need some advice on my first interactive perspective page..
With multiple users accessing the perspective page a set of dataset tags that sit on the gateway will be continually overwritten and on refresh users would potentially see data that is not correct to their own query.
AI suggests abandoning tags and pushing the datasets (pydatasets) directly to the tables, which sounds like a good theory? Client tags are Vision based but then session properties may be an option?
I do not want to build something and then discover I should have done it differently so I am reaching out for guidance.
Any opinions are welcome.
Thanks.