Add feature extraction with feature specifications

Adds a feature-extraction output that computes person-level summary columns (boolean, count, mean, etc.) from event-level data in a single OMOP table. Each omop_feature_spec in specs produces one column in the resulting data frame. Each spec is evaluated against its own concept scope. No output-wide concept prefilter is added: such a prefilter would change the meaning of an unscoped spec or of specs that use different concept_col values.

Usage

ds.omop.plan.features(
  plan,
  name,
  table,
  specs,
  grain = "person",
  temporal = NULL
)

Arguments

plan

An omop_plan object.

name

Character; output name used as a key in the plan's outputs list.

table

Character; source OMOP table name (e.g. "condition_occurrence", "measurement").

specs

Named list of omop_feature_spec objects created by the omop.feature.* family of functions (e.g. omop.feature.boolean, omop.feature.count).

grain

Character; "person" (default) or "episode". Episode grain preserves one row per cohort episode and requires temporal$index_window.

temporal

Optional omop.temporal() specification. Supply an index_window when grain = "episode" or when feature specs contain episode-relative time_window values.

Value

The modified omop_plan with the features output appended.

See also

omop.feature.boolean, omop.feature.count, ds.omop.plan.events

Examples

if (FALSE) { # \dontrun{
plan <- ds.omop.plan()
plan <- ds.omop.plan.features(plan,
  name = "lab_features",
  table = "measurement",
  specs = list(
    has_glucose = omop.feature.boolean(c(3004410)),
    glucose_mean = omop.feature.mean_value(c(3004410))
  )
)
} # }