Irregular time series

An irregular time series stores data for a sequence of arbitrary timepoints. Irregular time series are appropriate when the data arrives unpredictably, such as when the application records every stock trade or when electricity meters record random events such as low battery warnings or low voltage indicators. Irregular time series are also required for packed data, which includes hertz data and compressed numeric data.

Irregular time series store the time stamps for each element instead of storing offsets because the interval between each element can be a different length. Irregular elements persist until the next element by default and cannot be null. For example, if you query for the value of a stock price at noon but the last recorded trade was at 11:59 AM, the query returns the value of the price at 11:59 AM, because that value is the nearest value equal to or earlier than noon. However, you can also create a query to return null if the specified time stamp does not exactly match the time stamp of an element. For example, if you query for the price that a stock traded for at noon, but the stock did not have a trade at noon, the query returns a null value.

Hertz data and compressed numeric data are stored in irregular time series because multiple records are packed into each element. However, hertz and compressed time series require regularly spaced data and behave differently from standard irregular time series.