Compare Design and As-Built Survey Points

Confirm the roles and revisions of both sources

The design table represents intended positions; the measured table represents observations of the corresponding objects. Label those roles explicitly before importing. Two office versions of a design file are not automatically a design/as-built pair, and a file modified after adjustment may need an explanation of which observations it represents. Retain the source filenames, dates, revisions and any relevant control information so the interpretation can be reviewed independently.

Confirm horizontal coordinates, horizontal units, vertical units and height reference for both tables. Sharing a project name is not enough to establish compatibility. A missing or uncertain datum should be investigated rather than covered by increasing tolerances. The purpose of a tolerance is to express an acceptance rule for comparable quantities, not to conceal a reference mismatch. Test a documented control or a known unchanged point before evaluating the remainder of a batch.

Establish correspondence before calculating differences

Matching identifiers can provide a convenient starting point when the project convention is reliable, but identical text does not prove that two records describe the same physical object. Repeated IDs, renamed points or multiple measurements of a feature need explicit treatment. Keep unmatched records visible. A point with no design counterpart may be a legitimate additional observation, an omitted design record or a naming error; the software cannot choose that explanation on your behalf.

If identifiers differ, confirm pairs from project evidence and inspect their source rows. Do not choose a nearby design point merely because it makes the deviation small. Once a pairing decision changes, recalculate and retain the new matching basis. The comparison report should make it possible to identify both source records for every evaluated pair. A list of point names without source context is insufficient when duplicate names exist in either table.

Read signed differences and horizontal magnitude together

The component differences are measured minus design. A positive Easting difference means the measured position is farther east in the agreed grid; a positive Northing difference means farther north. The horizontal deviation is the magnitude of those two components and has no side sign. The vertical difference keeps its sign: positive means measured height above the corresponding design height under the shared height reference.

A synthetic example with components plus 0.03 east and plus 0.04 north has horizontal magnitude 0.05. If the height difference is plus 0.01, the report contains three component differences and one horizontal magnitude, not a single interchangeable error number. A recurring direction across many pairs may reveal a setup or reference issue even if individual magnitudes are small. Review the pattern and the source context rather than relying exclusively on a count of green rows.

Apply separate and explicit tolerance rules

Supply the horizontal and vertical tolerances required for the task. This tool compares horizontal magnitude with the horizontal threshold, and absolute vertical difference with the vertical threshold. Equality belongs to the within-tolerance result. Thus the synthetic three-four-five case is within at horizontal tolerance 0.05 and vertical tolerance 0.01 when the height difference equals 0.01. These are exercise values, not a general engineering acceptance recommendation.

A threshold of zero can be meaningful for an exact reference exercise, but it is not a substitute for defining a real project's permissible discrepancy and measurement context. Changing display precision should not change the classification of the underlying values. Preserve the actual thresholds and units in the report so a recipient can reproduce the decision. If the project evaluates axis components, three-dimensional displacement or another statistic, do not assume the horizontal-magnitude rule is an equivalent substitute.

Keep missing heights and unmatched points in the result

When both horizontal coordinates are valid but one height is absent, the pair can have a horizontal assessment and an unavailable vertical assessment. Its overall state is partial unless an evaluated component already exceeds a threshold. It must not become a complete pass by treating an absent height as zero. Similarly, an unpaired source row is not an evaluated point with zero deviation. Keep these categories distinct in the totals and in any downstream summary.

The supplied practice pair contains A within both boundaries, B exceeding the chosen thresholds, C with missing design height and one unmatched row on each side. The intended output is three paired results plus two unmatched source records. Compare that accounting with the four source records in each file. Counting only pairs can hide records that were never evaluated, while counting each pair twice can make a result appear more complete than it is.

Choose a different comparison when the reference object changes

If the measured location has no corresponding design point but a verified design TIN covers it, use point-to-surface comparison. That tool interpolates the design height from the original supported triangle at the measured plan location. It is not equivalent to pairing the observation with the nearest design vertex. Coverage gaps and unresolved overlapping geometry remain explicit rather than being replaced with a convenient nearest height.

Use file revision comparison when the question is what changed between two versions of a point list. Use common-control fitting only when a documented local relationship must be estimated, with independent check points and suitable coverage. Fitting a transform solely to make design deviations small would change the question being asked. A high or low deviation can help interpret a construction condition, but a pointwise comparison does not calculate an earthwork volume.

Deliver a report with enough context to review

Save the untouched sources, interpretations, reference description, matching decisions, tolerances and full result. Include the signed components, magnitudes and exception categories rather than only a screenshot of the summary. Inspect outliers and a few within-threshold examples against their source rows. A report table may be paginated for readability, while the complete export remains the appropriate deliverable for reconciliation and independent review.

Before reporting a suspected calculation defect, reproduce it with a minimal synthetic case, an independent arithmetic check and the same settings. State whether the unexpected behavior concerns pairing, units, numerical values or classification. Do not attach private project coordinates unless an appropriate exchange has been agreed. The browser provides a traceable comparison under your assumptions; the responsible project process still determines whether those assumptions, observations and acceptance criteria are valid for the physical work.

Questions and answers