Supported aggregation rules

Each rule defines behaviour in both directions of the instant grid, so aggregation and disaggregation are one operation (the geoscales split of extensive vs intensive quantities):

Usage

CALENDAR_RULES

Format

A character vector of length 7.

Details

weighted_mean

Down: copy. Up: mean weighted by the declared leaves$share of each source timeslice. For intensive quantities (load, price, efficiency). The default.

sum

Down: split equally across the timeslice's grid instants. Up: sum. Conserves totals; for extensive quantities (energy, cost).

mean

Down: copy. Up: plain mean over instants — a time-weighted mean, since the grid is uniform. Differs from weighted_mean exactly when declared shares differ from real-time coverage.

copy

Down: copy. Up: the common value, erroring if it is not constant. For timeslice-invariant scalars.

sd

Up: standard deviation over instants — the spread of the fine signal within each target timeslice. Aggregation-only; going finer it degenerates to 0/NA.

share

Share within parent: each source timeslice's value divided by the total over its parent timeslice. Unlike every other rule the result stays keyed by the source calendar's timeslices – to (or parent=) names the parent calendar or timeframe – so it cannot be mixed with other rules in one call. For building profile shares and distribution keys; requires the source timeslices to nest within the parent's. The two share rules mix freely with each other, being one computation.

logshare

The same computation as share – the values ARE shares – but figures draw it on a fixed log10 percent scale (0.01%..100%), where share gets a fixed linear 0..1 scale. Use it when sibling counts differ by orders of magnitude and the linear scale flattens the crowded groups.

Examples

CALENDAR_RULES
#> [1] "weighted_mean" "sum"           "mean"          "copy"         
#> [5] "sd"            "share"         "logshare"