Professional Detailing2 min read

Coating Trial Data Joins: Check That Product Codes Do Not Multiply Rows

Combining a results table with a product table can accidentally duplicate observations if the matching key is not unique. Check row counts and identifier relationships before trusting the merged coating-trial analysis.

Updated · NANOLAB Knowledge

The key distinction

A product code may appear several times because of packaging variants or document revisions. Joining on that code alone can attach one measurement to several records. The dataset then looks larger without containing more independent evidence, and averages or totals can become distorted.

Practical checks

  1. Define the intended relationship between the two tables.
  2. Verify that the matching key is unique where the join requires it.
  3. Reconcile row counts and unmatched records after merging.

Example

Illustrative example: a results file has one row per panel, while a product table has two packaging records for each formula. A join on formula code doubles the measurement rows. The correction is to use the appropriate identity or resolve the relationship, not to accept the larger sample count.

Scope and limits

Do not delete duplicated-looking rows without checking why they appeared. This guide concerns data preparation and does not prescribe a software tool. Preserve the original tables and the merge logic so the final analysis can be reproduced and traced to the actual observations.

Related guidance

Ask NANOLAB about the exact product and intended use. Use the current product documentation for formulation-specific instructions.

Professional Detailing2 min read

Coating Trial Data: Separate Zero, Missing and Not Applicable

A recorded zero, a missing observation and a not-applicable field have different meanings. Keep them distinct so that averages, rates and completion checks do not silently treat absent information as a measured outcome.