All Collections
Administrative
Tables and Datasets
Datasets - How to Build a Dataset
Datasets - How to Build a Dataset

A step-by-step guide to building a Dataset.

Greg Olsen avatar
Written by Greg Olsen
Updated over a week ago

A Dataset is created from one or many Tables and is used to collect and transform data for use in a Workbook. Before you create a Dataset, you will need at least one active Table. If you are referencing more than one Table in your Dataset, make sure you have correctly established connections between those Tables using the Numetric Dictionary. For an overview of the purpose of Datasets, see the Getting Started article Assembling Datasets.

To create a new Dataset, follow the steps outlined below:

1. Navigate to the Datasets section of Numetric, here all existing Datasets will be displayed and grouped according to their categories. Clicking on the title of an existing Dataset will allow you to view and edit the Dataset. 


To create a new Dataset, click the + New Dataset button in the top right corner.

2. Begin by naming your Dataset. Click on the New Dataset text at the top of the screen, and give your dataset a name and a category (used for Dataset organization).

3. Each dataset must contain at least one Table. Click the + Add a Table button to select the first table to add to your Dataset.

From the Table Selection screen, select the Table you would like to add.

4. The selected Table will now appear in the Dataset. You can click on the Table to select which columns of data you would like to include form that Table in your new Dataset.

 5. Each Dataset is required to have a minimum of one Table. However, if your selected Table contains a Relationship connecting it to another Table, you can add the related table into the Dataset.

To add a related table, click the +Join Table button. 

The Table Join section will automatically display any tables with a Dictionary Relationship to your existing Dataset. To add a related Table, simply click on the Table name.

If the joined table results in a one-to-one join, then you will not need to select any aggregation methods, you will simply be able to select which columns of data you would like to join to the Dataset. To do this, click the 0 Columns link located to the right of the Table name, then select the columns to join to the Dataset.

With a second table added to your Dataset, you now have the option to Join additional tables. These joined tables can be either be joined to the Primary Table in the Dataset, or they can be joined to the Joined Table.

6. In the event that any of your Joins result in a one-to-many join, you will need to select an aggregation method. 

For more information on what an Aggregation is, or how to determine which Aggregation method is best for your situation, see the Dataset Aggregations article.

To add an aggregation, click the 0 Aggregations link located to the right of the Table name. 

This will open all existing Aggregations applied to this table. To add an Aggregation, click the + Add Aggregation button.

Select the column you would like to aggregate, then select the aggregation method from the dropdown list.

Note: in one-to-many joins, only aggregated data will appear in the Dataset. Any column from the joined Table that does not have a defined Aggregation Method will not be joined to the Dataset.

7. As with Tables, you have the ability to modify the Column Properties of your Dataset. Click on the Column Properties tab at the top of the screen to modify your Dataset column properties.

 8. Dataset Transformations function exactly the same as Table Transformations. The main difference with Dataset Transformations is they allow you to conduct transformations with data from various Tables.
Dataset Transformations are available under the Transformations heading.

For more information on the difference between Dataset and Table Transformations, see the Assembling Datasets Article. For a detailed step-by-step of the Transformation process, see the Transformations articles.

9. To preview your Dataset, click on the Dataset Preview tab. Note - your Dataset must be published in order for the Dataset Preview to display.

10. With your tables joined, columns selected, and aggregation methods identified, you are ready to save your Dataset. Click the Publish button located in the top right corner of the screen to save and publish your new Dataset.

11. With your Dataset published, it can be shared with others in your organization. To share your Dataset, click the Manage Users icon, located to the left of the Publish button. Here you can enter the people with whom you would like to share the Dataset. 

You can also select between View and Edit access by clicking on the dropdown to the right of the user name.

Note: Clicking Publish will queue up your Dataset for indexing. This process may take several hours, to several days. Indexing all of your data prior to the creation of a Workbook allows Numetric to support lightning-fast reporting. With this indexing approach you can search, filter, adjust, and create new Metrics in Workbooks without having to re-index, or update your database.

If you wish to undo your changes, click the Revert button in the top corner. This will undo any changes made to the Dataset, and revert it back to the previous state.

 

Did this answer your question?