# The 4 value types in PI

When dealing with process data it is important to know what type a value is of. In PI we distinguish between four value types: constants, "interval" values with regular intervals, "raw" and "single" values with irregular date intervals. Depending on its type value trends in all charts are drawn as shown in the image.

`Constants`are values being valid in every moment in time like surface areas, physical and mathematical constants, given formula parameters and so on.

`Interval values`are given within regular intervals like minutes, hours and days. As average values they represent a process parameter within the given interval. Interval values are most common when analysing process data, especially 1 minute-values or 1 hour-values. Interval values require a defined set point as starting point for the interval, normally the full minute (0 secs), hour (0:00), day (00:00:00) and so on.

`Raw values`are values recorded only when changing. If a specific value does not change for one day nothing will be recorded for one day; if it changes every second every second a record takes place. Therefore raw values are valid for a period

__after__their timestamp. This compared to regular interval values which are valid for a period

__before__their timestamp.

`Single values`are often used to log events like reaching a threshold value or when an action took place. That means single values are valid not for a period but only for the moment in time, not before and not after this moment.

When importing values into PI's database the interval of each channel is checked a) based on the first rows of a value table before importing all values and b) while importing all values. These checks are crucial to ensure high level data integrity.

To be as flexible as possible many different date interval conversions are available in PI like changing the interval, counting the number of irregular occurring events on a regular interval basis or even sum up all values on a regular basis. Such features are not available in Excel & Co. "by default" but should be in a decent process data analysing tool.

Dr.-Ing. Martin Horeni, 05/2020