Shape · Matching and dedupe

Survivorship

Merges each group of duplicate records into one best record, picking each field by a rule such as most recent or most frequent.

Also known as: Golden record, master data

How it works

Key columns group the rows, usually the cluster_id from Match Grouping. Every other column is resolved by the Default rule: Most frequent, Most recent (by date), Oldest (by date), Maximum or Minimum; the two date rules need a Date column. Per-column rules override the default for named columns. NULLs are skipped, so a field wins from whichever row has a value, and a group where every value is NULL stays NULL; the date rules also skip rows with a NULL date. Ties go to the smallest value. The output has one row per key in the input's column order, sorted by key.

Inputs and outputs

Settings

{
  keyColumns: string[]
  defaultRule: "most_frequent"|"most_recent"|"oldest"|"max"|"min"
  dateColumn?: string
  overrides?: [{ column: string, rule: <same> }]
}

Example

Merge the matched contacts: key columns cluster_id, default rule Most recent with date column updated_at, and a per-column rule Most frequent for last_name.

More in Matching and dedupe

• Flag Duplicates — Adds a true/false column marking rows that repeat on the key columns, and keeps every row for review. A fuzzy mode catches near-duplicates such as Jon and John.
• Blocking — Finds candidate pairs of records that share a key, such as the same postcode, within one table or across two. It narrows down the pairs before fuzzy matching.
• Match Grouping — Turns matched pairs into groups: if A matches B and B matches C, all three share one cluster ID. Wire in the original records too and every record gets its cluster ID, with unmatched records in a cluster of their own.