Overwrite: Merge All Values

The Merge All Values overwrite operation collects the target field's value from every record in the duplicate group and combines them into a single delimited string in the golden record. Rather than choosing one winner, it preserves all the information from every contributing record.

Example showing four records with tag field values merged into a pipe-separated list in the golden record

How It Works

When Merge All Values runs for a field:

  1. It collects the field value from every record in the group (or from the data sources you have selected in the rule configuration).
  2. Null and empty values are excluded.
  3. Duplicate values are deduplicated — each unique value appears only once.
  4. The remaining unique values are joined with a pipe character (|) as the delimiter.
  5. The resulting concatenated string is written to the golden record.

Example: A group has three records with a Tags field containing: "VIP", "Prospect", and "VIP". Merge All Values produces VIP|Prospect — duplicates removed, values joined.

When to Use Merge All Values

Merge All Values is most useful for additive fields where all source values are meaningful and should be preserved:

  • Tags and categories — each source may have classified the record differently; keep all classifications.
  • Notes and comments — accumulate remarks from multiple systems into a single field.
  • Alternative names or aliases — preserve all known names for an entity.
  • Phone number lists — combine all known numbers when the target system supports multi-value fields.

Tip

Merge All Values is a good choice when you cannot afford to lose data during deduplication and your downstream system can handle multi-value strings or you plan to parse the delimited result later.

Data Source Filtering

In the rule configuration dialog, you can restrict which data sources contribute values to the merge. For example, if you only want tags from two out of three sources, select only those sources. Records from unselected sources are excluded from the combined output.

Warnings and Considerations

Field Length Limits

If your target database or downstream system has a maximum field length (e.g., VARCHAR(255)), a merged multi-value string can easily exceed it. Verify the target column definition before using Merge All Values on fields with many contributing records.

  • Deduplication is exact-match only — "New York" and "new york" are treated as different values. Standardize casing in Data Cleansing before merging to avoid duplicates in the output.
  • Order is not guaranteed — the values are combined in record order within the group, which follows the original data source load order. Do not rely on a specific sequence in the merged string.
  • Null handling — if all records in the group have a null value for the target field, the golden record field will also be null.

Combining with Conditions

Merge All Values can be paired with overwrite conditions like No Condition (always merge) or Field Is Empty (only merge if the master field is currently empty). The most common pairing is No Condition — you typically want to merge values unconditionally when using this operation.