Skip to contents

get_column_type_drift

Source: R/redcap_projects_history.R
get_column_type_drift.Rd

Compares the column definitions of a source table and a target table via information_schema.columns and returns the columns that are present in both tables but whose target type differs from the type the target should have (see desired_column_type). Used to detect when REDCap's own DDL revisions to redcap_projects (e.g. a widened VARCHAR) have drifted out from under the frozen redcap_projects_history mirror table, and to migrate any ENUM column in the target to VARCHAR.

Usage

get_column_type_drift(source_conn, source_table, target_conn, target_table)

Arguments

source_conn,

a DBI connection to the database containing source_table

source_table,

the name of the table to treat as the source of truth for column definitions

target_conn,

a DBI connection to the database containing target_table

target_table,

the name of the table to check for drifted column definitions

Value

a dataframe with one row per drifted column: column_name, source_column_type, target_column_type, desired_column_type, target_is_nullable

Details

Nullability is intentionally not compared: redcap_projects_history accumulates rows over time, so older rows may legitimately hold NULL in a column that REDCap has since made NOT NULL. Tightening the target to match the source's nullability would fail against that pre-existing data, so target_is_nullable is returned only so callers can preserve the target's own current nullability when reconciling type.

Examples

if (FALSE) { # \dontrun{
rc_conn <- redcapcustodian::connect_to_redcap_db()
rcc_billing_conn <- connect_to_rcc_billing_db()

drift <- get_column_type_drift(
  source_conn = rc_conn,
  source_table = "redcap_projects",
  target_conn = rcc_billing_conn,
  target_table = "redcap_projects_history"
)
} # }