Build the LP row index — build_row

One row per LP constraint. Rows are enumerated per equation from the tuples of the mapping named by equation$domain; an equation with no domain is treated as a scalar and contributes a single row.

Usage

build_row_index(model, verbose = FALSE, names = TRUE)

Arguments

model: A multimod model with data attached.
verbose: Logical; report per-equation row counts.
names: Logical; add the readable name column — see build_col_index().

Value

A data.table with columns i (1-based row index), symbol, name (when names = TRUE), i1..iN, sense ("==", "<=", ">=") and rhs (NA_real_ until coefficient extraction fills it).

Details

Counting is per equation, which is what an LP needs. energyRt's model_size() counts per map and therefore undercounts whenever one mapping gates several equations.

Examples

rows <- build_row_index(example_models$toy)
table(rows$symbol)
#> 
#> eqCapLim   eqCost eqDemand 
#>        6        1        2