We are going to create fake "metadata" associated with our 1000 time series, so that we can test a workflow in which we query timeseries by their metadata in a relational database, get results, and look for similar other time series.
(1) keeping the ids of the 1000 timeseries from previous project missives, add random metadata for a floating point field blarg which takes continuous values on [0,1] drawn from a Uniform distribution.
(2) randomly drawn from letters A-F to create a field level associated with each id. This is kinda like throwing a dice: there are discrete enumerated values A/B/C/D/E/F that can result
Metadata has been created and tested in a local sqlite database.
Timeseries are accessed through Storage Manager.
Sqlite database has been converted to Postgres database and the script was run on the EC2 instance.
We are going to create fake "metadata" associated with our 1000 time series, so that we can test a workflow in which we query timeseries by their metadata in a relational database, get results, and look for similar other time series.
(1) keeping the ids of the 1000 timeseries from previous project missives, add random metadata for a floating point field blarg which takes continuous values on [0,1] drawn from a Uniform distribution.
(2) randomly drawn from letters A-F to create a field level associated with each id. This is kinda like throwing a dice: there are discrete enumerated values A/B/C/D/E/F that can result