Not every tag in a historian comes from a scanning collector polling a live device. Some values are typed in by a technician or imported from a laboratory - a crude assay, a manual tank gauge, a water-cut sample. These lab and manual entry tags arrive rarely, at irregular times, and often out of order, and treating them like a scanned process tag produces nonsense. This guide explains what these sparse tags are and how historians handle them.
Lab & Manual Entry Tags in one line: A lab or manual entry tag is a historian point whose values are entered by hand or imported from an offline source rather than scanned from a live device. Values arrive sparsely and irregularly - a lab result hours after the sample, a manual gauge once a shift - so historians treat them as sparse, source-timestamped tags that must not be interpolated or compressed like continuous process signals.
A scanned process tag has a rhythm. The collector reads it every second or every minute, so the historian knows roughly when the next point is due and can flag the tag as stale if an update is late. A lab tag has no such rhythm. A crude oil assay might be logged once a day, a sediment-and-water result once a week, a manual tank strap once a shift. There is no scan interval, so the concept of a value being overdue does not apply the same way, and staleness checks tuned for a scanned signal will misfire.
These tags are also naturally out of order relative to the clock. A sample is drawn at 8:00 but the laboratory result is not known until noon, so the value is entered at noon carrying an 8:00 sample timestamp. That is a legitimate historical insert, not an error, and the historian must accept the entry against the sample time rather than the entry time. Getting this wrong shifts a lab result four hours into the future and misaligns it with the process conditions it was meant to describe.
Because the values are so far apart, the reconstruction rules differ. A continuous process tag can be linearly interpolated between compressed points because the underlying signal really did vary smoothly. A lab value describes a single instant - the moment the sample was taken - and says nothing about the hours before or after it. Drawing a straight line between two weekly assays implies a smooth trend that was never measured, so lab tags are usually held as the last known value or shown as discrete points rather than interpolated.
The single most important attribute of a manual or lab tag is the distinction between when the sample was taken and when the value was recorded. The historian should store the sample or observation time as the tag timestamp, because that is when the value was physically true. The entry time and the person who entered it are audit metadata - useful for traceability, wrong to use as the timestamp. Confusing the two is the classic manual-entry mistake.
Manual entry also introduces the possibility of correction. A technician may realize a gauge was misread or a lab may reissue a result, so historians that support manual tags usually allow an existing value to be edited or superseded, with the change tracked. This is a form of controlled out-of-order insert into the past, and it needs the same care as any backfill: any calculated tag or allocation figure that already consumed the old value should be reconsidered once it changes.
Because these values are entered by people, quality and validation matter more than for scanned data. A fat-fingered decimal point on a manual tank gauge can be a hundredfold error, and unlike a sensor fault there is no bad-quality flag to catch it. Reasonable systems bound-check manual entries against expected ranges and record who entered each value, so an obviously impossible number is questioned before it contaminates a report.
Oil and gas operations run on a mix of live telemetry and offline sampling. Continuous tags track pressure, temperature, and flow every few seconds, but the numbers that settle a custody dispute or an allocation - crude quality, water cut, gas composition, tank strapping - often come from a laboratory or a manual reading. A SCADA historian that only understood scanned tags could not hold this half of the record, so blending sparse manual tags alongside dense process tags is essential to a complete operational picture.
The value of combining them is context. A water-cut assay entered against its sample time can be overlaid on the continuous separator level and flow trends from the same period, letting an engineer connect an offline result to the live conditions that produced it. That only works if the lab value sits at its true sample time, not the time someone happened to key it in, and if it is not smeared across the gap between samples by interpolation.
In a cloud SCADA platform like Merobix, lab and manual entry tags live in the same historian as the scanned telemetry but are treated as the sparse, source-timestamped points they are. A field technician or lab tech can enter a value against the time the sample was taken, corrections are tracked, and the value shows as a discrete point rather than a fabricated trend - so the manual record and the live telemetry tell one consistent story.
A normal process tag is scanned from a live device on a fixed interval, so it produces a dense, regular stream of values. A lab or manual entry tag is entered by hand or imported from an offline source, so its values are sparse, irregular, and often recorded well after the sample was taken. Because of that, historians timestamp lab tags to the sample time, avoid interpolating between them, and do not apply the same staleness or scan-rate logic used for continuous signals.
It should be timestamped to when the sample was taken, because that is the moment the value was physically true. The entry time and the person who entered it are recorded as audit information, not used as the tag timestamp. Using the entry time shifts the result into the future and misaligns it with the process conditions it was meant to describe, which is a common and damaging manual-entry error.
Generally you should not. A lab value describes a single instant - the moment the sample was drawn - and says nothing certain about the time before or after it. Drawing a straight line between two sparse assays implies a smooth trend that was never measured. Historians usually hold a lab value as the last known reading or display it as a discrete point rather than interpolating a fictitious path between samples.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.