Payroll discrepancy case
Owner: Elmbrook Operations payroll reviewer.
Question: Which matched workers have different planned and paid hours?
Data: Fictional roster and payroll with duplicates, padding and unmatched source records.
Done: A matched discrepancy table with Hours, Paid and Variance; unmatched streams remain separate evidence.
Elmbrook Operations sees a difference between planned and paid hours. A balanced overall amount would not prove that every worker matched correctly, because positive and negative differences can cancel. The reviewer asks for the matched workers whose planned hours differ from payroll, along with the source values needed to explain each gap.
Prepare worker identifiers and remove the repeated HR row before matching. Keep the workflow's L and R outputs available to inspect source-only workers; those are a separate population from the matched discrepancies in the requested final table. Calculate a signed difference and then filter for nonzero gaps. This distinction prevents an unmatched worker from being described as a paid-hours variance. The hidden payroll period changes which matched workers need review.
input: {"source":"main"}
formula: {"field":"Worker","expression":"Trim([Worker])","type":"V_WString"}
Worked example: A padded key must be normalized before classifying it as missing.