Skip to content

Handle missing result identifiers without accidental merges

Separate unkeyed rows and prevent null or empty values from matching each other.

Guides

Handle missing result identifiers without accidental merges — what should the operator record?

Several returned rows may lack the field your transformation expected to use as a key. Joining all missing identifiers together can create an invented relationship among unrelated evidence. Quarantine unkeyed rows, preserve their receipt references and decide whether another actual source field can support a justified identity contract.

Practical operations check
Separate unkeyed rows and prevent null or empty values from matching each other.
Distinguish absent key, explicit null, empty string and parse failure in an identity_state column rather than collapsing them into one value. Whitespace-only values need an explicit contract too: do not automatically strip characters from an opaque key unless the actual source representation justifies that normalization.
Keep actual accepted counts separate and state which records could not participate in joins, without dropping them silently from the handoff. The handoff can include keyed_rows and unkeyed_rows as separate artifacts, with actual parsed membership and a clear blocker for operations that require stable identity.
01 / Guides

Classify missingness

Distinguish absent key, explicit null, empty string and parse failure in an identity_state column rather than collapsing them into one value. Whitespace-only values need an explicit contract too: do not automatically strip characters from an opaque key unless the actual source representation justifies that normalization.

02 / Guides

Block null equality joins

Require valid qualified identifiers on both sides before matching; missing keys should enter an exception table rather than match other missing keys. In common data tools, null join behavior varies; verify the selected operation with fixtures rather than assuming missing values never match or always match everywhere.

03 / Guides

Assign local references carefully

Use a receipt-and-row-position reference for operational review only, labeling it nonpersistent and unsuitable for cross-retrieval identity. A receipt row reference can help an operator locate evidence within that exact artifact but should not be reused as identity across reordering or new retrievals.

04 / Guides

Review alternative evidence

If an original canonical reference is available, assess it under a documented rule instead of guessing identity from title or author similarity. A canonical URL may support a local key only if its identity semantics are understood; removing query components indiscriminately can collapse distinct source objects.

05 / Guides

Release keyed and unkeyed views

Keep actual accepted counts separate and state which records could not participate in joins, without dropping them silently from the handoff. The handoff can include keyed_rows and unkeyed_rows as separate artifacts, with actual parsed membership and a clear blocker for operations that require stable identity.

Keep the boundaries clear.

  • Separate unkeyed rows and prevent null or empty values from matching each other.
  • Operational record: Classify missingness. Context badges do not certify access.

Start with a read-only check: ask your agent to report actual source coverage before collecting records or exporting an audience.

Primary references

Documentation reviewed 2026-11-15. Product capabilities, client availability and plans can change. Third-party marks identify their products, not an affiliation.

Know what social media shows about you.
Start with a question.

Get started
montage