Looking for a TSBD (Time Series Database) that supports multiple observations for the same Timestamp

user_19740958 0
Viewed 36

I am looking for a time series database that supports storing multiple values for each observation.

To be precise, it is about weather forecasts that change multiple times and thus multiple values exist for each TimeStamp. Therefore, multiple values must be stored for each TimeStamp, which themselves have a timestamp (namely the time at which the forecast was created).

Does anyone know of a TSDB that supports such functionality? Or do I have to assemble something myself?

Many thanks and best regards

Johannes

time-series database
1 Answers

Disclaimer: I'm working for TimescaleDB

TimescaleDb (timescaledb.com) is an extension built upon Postgresql. That said, you can use all the existing functionality and analytical tools from Postgres itself. Upon that, Timescaledb adds an extension to the query planner which makes sure that time-series data (stored in Timescale's Hypertables) are stored in time-based chunks. That optimizes the ingest and query speeds.

In terms of Hypertable design, you have multiple options. Hypertables are an extension of the standard database tables and can have multiple value columns per row, but you can also store the values one-per-row.

If you have multiple rows per timestamp and location (or however you have grouped your observations), you can use standard SQL (such as max) or advanced Timescale-provided functions (such as last) to retrieve the latest forecast.

user_9173360
Related