How to pass multiple options to pg_dump?

[janklan]

I'm trying to exclude several tables from the dump. Normally this would be done by pg_dump --exclude-table-data "pattern1" --exclude-table-data "pattern2", etc ...

YAML doesn't allow duplicate keys in a map:

dump:
    pg_dump_options: # pg_dump option that will be provided
        exclude-table-data: "pattern1"
        exclude-table-data: "pattern2"

pg_dump_options can't be turned into an array:

dump:
    pg_dump_options: # pg_dump option that will be provided
        - exclude-table-data: "pattern1"
        - exclude-table-data: "pattern2"

and passing a regex isn't supported by pg_dump:

dump:
    pg_dump_options: # pg_dump option that will be provided
        exclude-table-data: "(pattern1|pattern2)"

---

How do I exclude data from multiple tables in different schemas?

[wwoytenko]

Hi! This parameter is a list, you can set a list of tables to exclude.

exclude-table-data: ["tab1", "tab2"]

[wwoytenko]

Regexp should has been working, if not so, please open a bug report

[wwoytenko]

Tested in the playground:

Here is a config

dump:
  pg_dump_options: # pg_dump option that will be provided
    dbname: "host=playground-db user=postgres password=example dbname=original"
    jobs: 10
    exclude-table-data:
      - "addr*"
      - "businessentity"

Here is a debug output:

2025-07-07T06:37:35Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=businessentity pid=76
2025-07-07T06:37:35Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=address_addressid_seq pid=76
2025-07-07T06:37:35Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=address pid=76
2025-07-07T06:37:35Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=addresstype_addresstypeid_seq pid=76
2025-07-07T06:37:35Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=addresstype pid=76

I checked the tables - they are empty.

The "or"-like regexp works fine as well

dump:
  pg_dump_options: # pg_dump option that will be provided
    dbname: "host=playground-db user=postgres password=example dbname=original"
    jobs: 10
    exclude-table-data:
      - "(addr*|businessentity)"

2025-07-07T06:42:06Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=businessentity pid=28
2025-07-07T06:42:06Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=address_addressid_seq pid=28
2025-07-07T06:42:06Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=address pid=28
2025-07-07T06:42:06Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=addresstype_addresstypeid_seq pid=28
2025-07-07T06:42:06Z DBG ../greenmask/internal/db/postgres/context/pg_catalog.go:104 > object data excluded TableName=person TableSchema=addresstype pid=28

The only thing that I've found is that it's required to switch TableName and TableSchemain the log - it should beTableName=addresstype TableSchema=person

[janklan]

Hi! This parameter is a list, you can set a list of tables to exclude.

exclude-table-data: ["tab1", "tab2"]

Thank you! I (1) didn't know it's a list and (2) probably also messed up the regex.

All sorted now.