// Template · Governance

RDBMS Impact + SQL Parser

A column change impact report that catches every database reference across your entire Hop project — both explicit metadata references and table names buried inside SQL text. Run it before any schema change.

Uses RDBMS Impact SQL Parser

Governance Data Quality Impact Analysis Schema Changes Putki plugins

Get access

Apache Hop pipeline (.hpl)
Available to Putki subscribers

Pipeline structure

get-hop-files

GetFileNames

Putki

scan-hop-project

RDBMS Impact

has-explicit-table

FilterRows

A · metadata

select-metadata-fields

SelectValues

add-source-metadata

Constant · METADATA

Append streams

Append

write-impact-report

TableOutput

B · sql text

select-sql-fields

SelectValues

Putki

parse-transform-sql

SQL Parser

filter-empty-parsed

FilterRows

add-source-sql

Constant · SQL

How it works

RDBMS Impact scans the project and catches explicit table references — transforms where the table is configured directly in the metadata. But it misses table names buried inside SQL text. SQL Parser handles exactly that: it parses the SQL statically and extracts every schema, table, and column referenced. Together they cover the complete surface area of your project's database dependencies.

Requires RDBMS Impact and SQL Parser. Both are Putki plugins — not available in the open-source Apache Hop distribution. A Putki subscription is required to run this template.

List all pipeline and workflow files

get-hop-files scans ${PROJECT_HOME} recursively and returns one row per .hpl and .hwf file found. Each file path is passed downstream to the RDBMS Impact scanner.

Scan each file for database references

scan-hop-project (RDBMS Impact) reads every transform and action in each file and emits one row per database reference found — table, connection, schema, item name, and the raw SQL text when available.

Split into Stream A and Stream B

has-explicit-table routes rows with a table name defined directly in metadata to Stream A. Rows where the table field is null — but the transform has embedded SQL — go to Stream B for parsing.

Stream A — tag explicit references

Fields are renamed for consistency (schema → schema_name, table → table_name, filename → object_name) and a source = METADATA tag is added to every row.

Stream B — parse the embedded SQL

parse-transform-sql (SQL Parser) reads the sql_text field and extracts every schema, table, and column referenced. It doesn't execute the SQL — it parses it statically. Rows where parsing returns no table are dropped by filter-empty-parsed. A source = SQL tag is added to the remaining rows.

Merge and write

Both streams are appended into a single output and written to public.column_change_impact. The table is truncated on every run, so the report always reflects the current state of the project.

Querying the report

Run this before renaming, dropping, or changing any column. An empty result means the change is safe. If it returns rows, you have a list of every pipeline, transform, and source type that will be affected.

SELECT object_name, item_name, item_type, source
FROM   public.column_change_impact
WHERE  table_name  = 'customers'
  AND  column_name = 'email'
ORDER BY source, object_name;

Output fields — column_change_impact

Field	Description	Source
connection	Name of the database connection in the Hop project metadata	Both
schema_name	Database schema (e.g. `public`), if specified	Both
table_name	The table or view being referenced	Both
column_name	Column referenced inside the SQL statement	SQL only
object_name	Full path of the `.hpl` or `.hwf` file containing the reference	Both
object_type	Whether the file is a Pipeline or a Workflow	Both
item_name	Name of the specific transform or action inside the pipeline	Both
item_type	Transform type — e.g. `TableInput`, `Execute SQL`	Both
sql_text	The raw SQL text found in the transform configuration	SQL only
source	METADATA for explicit table refs · SQL for parsed SQL text	Both

What you need

Putki subscription

RDBMS Impact and SQL Parser are Putki plugins. Both must be present in your distribution to run this pipeline.

A PostgreSQL connection named postgres

Configure a connection with this name in your Hop project metadata, or update write-impact-report to point to your connection and schema.

The output table created in advance

Run the DDL included in the download to create public.column_change_impact before the first execution. The pipeline truncates and rewrites it on every run.