One place, more than one flow

Two live flows carry the same substance, the same context and the same unit -- three of the fields deduplication signs on. They were not collapsed, because it compares every other field of the record as well and found one they disagree on. That field, shown as what holds them apart, is the whole of what keeps them two. Where it is the general comment -- a free-text note -- they are one edit to a source list away from being merged, and the flow holding fewer characterisation factors would lose them. That is #255, where a balanced pair of water withdrawal and return factors became a one-sided charge. Everything else the flows disagree on is listed under each of them: the names, most often, which deduplication does not read at all and a curator does. So: two flows, or one? Answering one means a ruling in `elementary-flow-collision-decisions.json`, which names the flow that survives and carries the other's factors and source references onto it. Answering two means leaving them alone, which is what the pipeline already does. A row marked `review` is a different question and takes a different answer: there the merge minted the flows, answering a curated grouping by writing a flow onto a place another flow already held, where the same grouping attaches a source reference to the flow that is there when the substance is in the prepared correspondence table (#361). Those minted flows carry no characterisation factors, and a ruling in the decisions file does nothing to them -- deduplication and the rulings both run before the merge, so nothing has ever inspected these. What answers them is which of the two treatments the grouping should have had.

elementary-flow-collision