View connection history

Available in Beta

For any data connection you’ve created in Pecan, you can now see a log of all table imports or prediction exports (depending on the type of connection), along with detailed information about each activity.

Select a start date for data imports

Available in Beta

Processing large amounts of unused data increases compute time and may generate additional costs.

By selecting a start date in a table’s import settings, you can control how much data is imported to Pecan – saving data-transfer costs and reducing the time required to train models and generate predictions.

To use this feature:

  1. Go to the ”Connections” screen and locate the relevant table.

  2. In the dropdown menu for the table, select Settings.

  3. Go to the “Advanced” tab.

  4. Under “Select start date”, select the relevant column (supported column types include date and timsetamp). Then select the date you wish to begin importing data from.

New Delta file connector

Available in Beta

Delta Lake is an open source storage framework that enables building a Lakehouse architecture. Delta tables are the underlying format and structure of files and folders that make up this framework.

Pecan is now able to import data from Delta tables so they can be used to train models and provide the basis for predictions. You can also export predictions from Pecan to Delta tables.

To connect to a hosted Delta file, go to the “Connections” screen, click Add connection, select Delta file, and complete the required fields.

Other product improvements

Available in Beta

In Pecan, you can change the way tables are automatically imported as part of a recurring prediction schedule. This feature has now moved into each table’s Settings menu.

To select or change the default import method:

  1. Go to the ”Connections” screen and locate the relevant table.

  2. In the dropdown menu for the table, select Select automatic import method.

  3. Choose your desired method and click Save.

The default or selected method will be labeled as “Default”.

Did this answer your question?