1

Hi all, I’m looking for some high level design feedback. I am working on a redash dashboard and I was hoping to have a dashboard-level filter that lets the user select what time period they want data from (i.e. only include data from the past 24 hours, or only include data from July 23-28, 2022, etc.) According to the redash docs on query parameters, dynamic dates and date ranges aren’t compatible with Scheduled Queries. So, I was thinking of having 2 sets of queries:

  1. One set of queries would query the datasource (hive, redshift, …) directly at a scheduled interval (once a day/week/month/ever).
  2. A second set of queries would use redash QRDS on the cached results of the first queries, with date range dashboard level filters

Is this a reasonable design? Would it take too long on the client-side? Would it be too computationally expensive? Any advice is appreciated

2

Welcome to the forum!

Yes that’s a reasonable design.

The time it takes on the client varies on how many rows there are. Keep in mind that Redash is a visualisation tool for aggregated data sets. It’s not meant to pull 1m or even 100k rows into your browser, although it can in some cases do that.

Give it a try and ping if you hit any blockers :grin: