With IBM® Informix® 12.10.xC5, you can search time series data for matches to a specific pattern of values. For example, after you identify a pattern of abnormal electricity usage that indicates an outage, you can search for matches to that pattern to find other outages. A pattern is a sequence of numeric values in a field within the TimeSeries subtype. You can specify a search pattern as a time range in a specific time series or as a list of numeric values. A match is a sequence of values from a target time series that conform to the search criteria.
When you run a pattern matching search, you specify the method to evaluate possible matches and the margin of error between the search pattern and matching values.
The following information is returned for each resulting match:
- The primary key value for the time series instance.
- The begin point of the match.
- The end point of the match.
- The score of the match, as a double precision number 0.0-1.0 that represents the percentage of values in the matching sequence that are an exact match to the corresponding values in the search pattern.
You can create a pattern matching search index on a time series instance to improve query performance.
Read more: Pattern matching searches.
Attend the Using Pattern Matching Analytics in IoT and TimeSeries data Webinar on April 20: https://events.na.collabserv.com/register.php.