Aglet

Triage Discrepancies in an Exported Dataset

An export can disagree with its source because the population moved, fields were transformed, rows were paginated, or the file reflects a different snapshot. Triage should define the export contract, capture source and file metadata, and compare identifiers and values in a bounded sample.

Establish what is happening

  1. Define source and file

    Record the source view, filters, ordering, selected fields, time zone, snapshot or generation time, file format, encoding, and expected row or byte boundaries. Note whether totals describe rows, entities, or groups. Preserve the exact request and file metadata.

  2. Compare records and fields

    Match a bounded set of source identifiers to exported rows and classify missing, extra, duplicated, reordered, renamed, truncated, formatted, or stale values. Include null, zero, long, and special-character examples. Keep raw source and file values side by side.

  3. Bound timing and impact

    Group differences by export run, source revision, field, entity type, filter, and generation time. Recalculate affected totals and downstream uses. Separate a moving source snapshot from deterministic serialization or selection behavior, and record the first boundary that diverges.

What to carry forward

Deliver an export scope brief with source and file contracts, matched identifiers, discrepancy classes, timing, affected runs, and evidence limits. Triage stops when the difference can be reproduced from recorded inputs. Keep the cause open when the source changed and no comparable snapshot exists.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow