Processing type
A Basic Metric is stored as an on-demand metric by default.
To change this setting, click the Change button at the top of the screen, above the field and record columns. Then select one of the following storage options:
-
On demand, do not store
-
Store, using Overwrite logic
-
Store, always append to existing
Overwrite vs. append to existing values
- Overwrite Existing Values
-
This setting is usually selected since it ensures that the latest metric value is stored, even when data for a given time period is loaded more than once. For example, you might load the metric data for a particular month and then reload it during the same month.
This may happen if the data is changed and you want to refresh the MetricsMart with metrics calculated from the new data, or if you change a metric or record definition and need to reload the data to get results based on the new definition. In either case, "Overwrite existing values" replaces the data previously loaded for the month with the new values.
You can create metric pre-aggregated tables (MPATs) for each metric to expedite calculation and processing data.
- Append to Existing Values
-
Select this setting if results are meant to be cumulative. This means that as new data is loaded, it is added to existing data, even if previous metrics for the same time interval are reloaded. This setting is often used if multiple data sources are contributing to a single metric. For example, if you have multiple data sources, after Data Source A is checked for the day’s data, you can check Data Source B for the day’s data and the data from both A and B are added together to get the final metric value for the day.
If you select either of the Store (Pre-Processed) settings, select your storage location and time options for frequency of overwriting or appending.
An On-Demand Metric is available immediately to a dashboard. The No Data Load step is required because the data is loaded and calculated as it is needed.
A Pre-Processed Metric pulls data on a schedule and then calculates and stores the results in a MetricsMart. The dashboard pulls the data from the MetricsMart.
If you convert a Basic Metric from On-Demand to Pre-Processed, you must complete a data load before the metric results will be available to a dashboard.
Calculation interval
The Calculation Interval sets the smallest unit of time for which you are able to view data.
If the interval is set to 1 day, one entry is made in the metric table for each day for each dimension combination.
This enables plotting data on a day-by-day basis. It also allows plotting data against any time interval larger than 1 day, such as by week or month.
A calculation interval of 1 day would not allow plotting a chart of data hour by hour. That would require a calculation interval of 1 hour or less. Before setting the granularity of the time interval, it is important to consider the number of entries to the MetricsMart that will be generated for each metric. The smaller the time interval you select, the more data there will be in your MetricsMart. Always select the smallest time interval you need to keep track of, but do not select a smaller interval than necessary.