There are three main ways to filter Topic content.
Content Filtered Topic – This is executed with the frequency of arriving data. If data is received at a high frequency but would be read (taken by the Data Reader) at a low frequency, then this might not be the most efficient approach. This option is fundamentally different from a Query Condition/Filter in that data is filtered before it is inserted into the Data Readers cache, therefore keeping the Data Readers cache relatively small, containing only relevant data.
Instance-Handle Filter during read – This option allows you to manually filter content based on the topic key. This is a very efficient way of obtaining data of a specific instance as it doesn’t require the execution of any SQL query on content. It instead ‘points’ directly to the key-value for which the data is requested.
Query Condition/Filter– In this case a query is typically executed with the frequency of a read/take so should be used if you read/take data at a low-frequency. This, for example, can be used with the KEEP_LAST history option for QoS to ‘down sample’ data for each instance.
shows more detail on these in the tutorial.