Filter Dedup


This filter removes duplicate records. A record is considered a duplicate, and is thus removed by this filter, if another record with the same values has already been seen. The comparison is performed on a user-provided list of fields (Fields setting).

WARNING: to remove duplicates, this filter stores one key per unique record in memory, this means that the overall memory grows linearly with the number of unique records in your data set. Depending on your data set, this might lead to OOM (i.e. out of memory) errors.


Keys available in the [filter.config] section:

Name Type Default Required Description
Fields array of strings [] true fields to consider when comparing records
KeySeparator string “\x1e” false character separator used to build a key from the fields
Last modified April 6, 2021