From oracle-ai-data-platform-workbench-engineer-agent
Profiles an AIDP table via Spark SQL — row count, per-column null %, distinct count, min/max/mean, and top-K values. Use for data-quality snapshots or understanding dataset shape.
How this skill is triggered — by the user, by Claude, or both
Slash command
/oracle-ai-data-platform-workbench-engineer-agent:aidp-profiling-tablesThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Produce a column-level profile of an AIDP table via Spark SQL. Self-contained: control-plane lookups
aidp-profiling-tables — single-table profileProduce a column-level profile of an AIDP table via Spark SQL. Self-contained: control-plane lookups
use oci raw-request; profiling SQL runs through the bundled scripts/aidp_sql.py helper. No aidp MCP
server is required.
aidp-catalog-explore / .aidp/catalog.md) → fully-qualified catalog.schema.table
and its columns/types. Without a cache, list via oci raw-request:
GET /tables?catalogKey=<cat>&schemaKey=<cat.schema> and filter for the table client-side (see
references/no-mcp-rest-map.md). Use the column types to pick the
right per-column profiling SQL.python "$PLUGIN_DIR/scripts/aidp_sql.py" --region <r> --datalake <ocid> --workspace <ws> --cluster <key> \
--code "spark.sql('''<profiling SQL>''').show(50, truncate=False)"
SELECT COUNT(*) FROM t (flag if LARGE; sample for the rest).MIN, MAX, AVG, COUNT, null %, approx distinct (approx_count_distinct).approx_count_distinct, top-K via GROUP BY … ORDER BY count DESC LIMIT k.MIN/MAX range, null %.
Use TABLESAMPLE/LIMIT on large tables to stay cheap; say when you sampled. The helper returns JSON
(status, outputs, spark_job_ids) — parse outputs for the result rows..aidp/catalog.md value dictionaries (aidp-catalog-init) and to add
data-quality rules (aidp-data-quality).--session-profile AIDP_SESSION only if your tenancy is session-token-only. On a kernel/auth error,
refresh (oci session refresh --profile AIDP_SESSION) and retry.aidp-data-quality, aidp-catalog-initnpx claudepluginhub anthropics/claude-plugins-official --plugin oracle-ai-data-platform-workbench-engineer-agent2plugins reuse this skill
First indexed Jun 12, 2026
Generates detailed profiles of database tables including metadata, row counts, column statistics, cardinality analysis, sample data, and quality checks for completeness, uniqueness, and freshness.
Profiles unfamiliar datasets: schema structure, column distributions, null rates, cardinality, outliers, table relationships, and temporal coverage. Onboard new data sources, audit freshness, or discover foreign keys.
Validates AIDP tables against data-quality rules (not-null, uniqueness, range/set, referential integrity, freshness) using bounded Spark SQL. Reports pass/fail with violation counts and can persist rule sets for re-runs.