Aglet

Investigate Source and Export Data Differences

Investigation should compare an export with the same source population under the same time boundary. Freeze or bound inputs, match identifiers, and trace selected fields through transformation and serialization. Test moving-source, query, mapping, and format hypotheses separately so the brief supports a specific correction.

Build a useful investigation brief

  1. Freeze the comparison

    Choose a source snapshot or exact query window and capture filters, ordering, selected fields, revision, and generation time. Obtain the corresponding file and checksum or byte metadata. If inputs moved, identify the changed records and state the comparison limitation before proceeding.

  2. Trace row and field paths

    Follow matched and unmatched identifiers through selection, joins, field mapping, formatting, escaping, encoding, line construction, and file writing. Compare raw and serialized values for nulls, delimiters, precision, long text, and special characters. Record intermediate counts and the code revision.

  3. Test one boundary

    Vary one filter, page boundary, timestamp, mapping, sort tie-breaker, or serialization rule at a time. Recalculate identifiers, values, and totals after each change. If source movement and file behavior both explain differences, separate the conclusions and name the missing snapshot evidence.

What to carry forward

Produce a reproducible export brief with fixed inputs, file metadata, matched sets, transformation traces, tested boundaries, and unresolved timing. Recommend the smallest correction or contract decision plus a bounded reconciliation cohort. Do not call a file accurate from matching totals when row identities differ.

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