Skip to contents

Validate a data frame against a schema's canonical column declarations

Source: R/crt_schema_validate.R
crt_schema_validate.Rd

Checks that every column declared required: true in the schema's canonical.cols is present in names(df). Fails loud listing all missing required columns. Returns invisible(NULL) on success.

Usage

crt_schema_validate(df, schema)

Arguments

df

A data frame (handler output, post-shape, pre-type).

schema

The parsed schema YAML (a list with canonical$cols).

Value

invisible(NULL) on success. Throws fail-loud cli::cli_abort listing all missing required columns on failure.

Details

Called by crt_ingest() AFTER handler dispatch and BEFORE crt_schema_apply() — validate shape first (fail loud on missing required cols), coerce types second. Without validation here, a missing required column would silently become NA after as.integer(NULL)-style coercion in the type-apply step; this surfaces the failure at the right layer.

Lives in the crt_schema_* family alongside crt_schema_apply() and crt_schema_read().

Future extensions (reserved family slots; NOT in v0.0.2):

  • cols[].range numeric range checks

  • cols[].enum value-membership checks

  • cols[].predicate custom predicate function refs

Each lands as additional cli::cli_abort branches in this function as the schema YAML grows the corresponding declaration keys. The function signature does not change — schema YAML drives behavior.