Introduction
A data key is a multi-field key intended for use in caching and similar scenarios.
Internals
Exact implementation details for data key may vary depending on language. However, in each case, the primary features are an integer hash code and a list of values in a specific order.
The hash code is calculated when the data key is created. It can subsequently be used as an efficient method to identify if one data key instance may match with another.
When a hash match occurs, each individual key field value should also be checked for equality between the keys to ensure a true match. This is necessary because the possibility space for a large number of keys with arbitrary value types is far larger than that of a 32-bit integer, so hash collisions can occur.
Usage example
A data key builder is used to construct data keys by specifying the details of the keys to be included. This will often be a subset of the evidence keys stored in a flow data instance. The diagram below shows an example data key being generated from the latitude and longitude values in an evidence collection.
The data key can then be used to quickly check if the current flow data contains the same values for the subset of evidence keys we're interested in as any other data key instance from a list.
First, the cache will look for a match on just hash code as this is a fast operation. In the example above, the generated data key matches two of the items in the cache on hash code. When the hashes are found to match, the cache will then compare the values of the key fields to check for a true match.