TRY ME

Try Valo for free

We want to show you something amazing.

We'll send you a link to download a fully functional Valo copy to play with.



Great! Check your email and enjoy Valo



Apologies, we seem to be having a problem processing your input, please try again

Time Series Repository

The Time Series Repository (TSR) is a custom built data repository designed to handle numerical time series oriented data streams. By time series we mean a series of data points each of which has an associated time stamp.

The TSR is also column oriented (unlike the SSR which is document oriented). Column oriented databases store each field in a specific file. This is helpful for certain operations as the DBMS does not need skip over other fields as with row oriented databases, or de-serialise whole document blobs as with document oriented databases. Since like data is ‘lined up’, it is possible to compress data better and to perform scan operations much faster. This is more suited to analytical operations.

The TSR is optimised to rapidly scan through time series data, and apply analytics to the data. All of the Scalar and Online Algorithm’s available to the Execution engine and to the SSR are also available within the TSR. This makes for a seamless experience in writing and deploying algorithms across all kinds of real time and historical data.

Configuration

The TSR involves very little configuration. Any stream to be stored in the TSR just requires the TSR repo mapping:

{
  "name" : "tsr"
}

More configuration will be introduced in time, including more control over indexing.

Indexing

By default, the TSR indexes all fields in a schema.

Note

This restriction will be lifted soon

Indexing allows for faster and more powerful query operations. While they are useful, they also have an overhead in both disk space and write speed.

All types within the Valo type system are supported within the TSR.