Configuration

DeepDiff DB is configured through a YAML file. By convention it is named deepdiffdb.config.yaml and lives in the working directory, but you can point any command at a different file with the --config flag.

Full Annotated Example

# ─────────────────────────────────────────────
# Production database connection (read-only)
# ─────────────────────────────────────────────
prod:
  driver: "postgres"       # mysql | postgres | postgresql | sqlite | mssql | oracle
  host: "127.0.0.1"
  port: 5432
  user: "postgres"
  password: "secret"
  database: "prod_db"

# ─────────────────────────────────────────────
# Development database connection (read-only)
# ─────────────────────────────────────────────
dev:
  driver: "postgres"
  host: "127.0.0.1"
  port: 5432
  user: "postgres"
  password: "secret"
  database: "dev_db"

# ─────────────────────────────────────────────
# Tables and columns to exclude from all diffs
# ─────────────────────────────────────────────
ignore:
  tables:
    - "audit_logs"         # exact table name
    - "sessions"
  columns:
    - "*.updated_at"       # wildcard: ignore updated_at on every table
    - "*.created_at"
    - "users.last_login"   # table-qualified: ignore only on users table

# ─────────────────────────────────────────────
# Where to write reports and migration files
# ─────────────────────────────────────────────
output:
  dir: "./diff-output"     # created automatically if it does not exist

# ─────────────────────────────────────────────
# Safety controls for schema migration
# All default to false. Set to true to allow.
# ─────────────────────────────────────────────
migration:
  allow_drop_column: false
  allow_drop_table: false
  allow_drop_index: false
  allow_drop_foreign_key: false
  allow_modify_primary_key: false

# ─────────────────────────────────────────────
# Conflict resolution
# ─────────────────────────────────────────────
conflict_resolution:
  # Default strategy for every table: ours | theirs | manual
  #   ours   — keep production values
  #   theirs — use development values
  #   manual — require interactive decision
  default_strategy: "manual"

  # Per-table overrides (applied before default_strategy)
  strategies:
    - table: "feature_flags"
      strategy: "theirs"   # always accept dev version of feature flags
    - table: "config"
      strategy: "ours"     # never overwrite production config rows

# ─────────────────────────────────────────────
# Performance tuning (v0.7+)
# ─────────────────────────────────────────────
performance:
  # Rows per keyset-paginated query during table hashing.
  # 0 = load entire table in one query (pre-v0.7 behaviour).
  hash_batch_size: 10000

  # Maximum tables hashed concurrently.
  max_parallel_tables: 1

Connection Options

Field	Required	Description
driver	Yes	Database driver identifier (see table below)
host	Yes (except SQLite)	Hostname or IP of the database server
port	No	Port number. Defaults: MySQL 3306, PostgreSQL 5432, MSSQL 1433, Oracle 1521
user	Yes (except SQLite)	Database username
password	Yes (except SQLite)	Database password
database	Yes	Database name. For Oracle: the service name (e.g. XEPDB1)

Driver Connection Snippets

MySQL

prod:
  driver: "mysql"
  host: "127.0.0.1"
  port: 3306
  user: "root"
  password: "secret"
  database: "prod_db"

PostgreSQL

prod:
  driver: "postgres"
  host: "127.0.0.1"
  port: 5432
  user: "postgres"
  password: "secret"
  database: "prod_db"

Both postgres and postgresql are accepted as the driver value.

SQLite

For SQLite the host, port, user, and password fields are ignored. Set database to the file path.

prod:
  driver: "sqlite"
  database: "/var/data/prod.db"

dev:
  driver: "sqlite"
  database: "/var/data/dev.db"

Microsoft SQL Server (v0.8+)

Port defaults to 1433 when omitted.

prod:
  driver: "mssql"
  host: "127.0.0.1"
  port: 1433
  user: "sa"
  password: "StrongP@ss1word!"
  database: "prod_db"

Oracle (v0.9+)

Port defaults to 1521 when omitted. Set database to the Oracle service name.

prod:
  driver: "oracle"
  host: "127.0.0.1"
  port: 1521
  user: "system"
  password: "secret"
  database: "XEPDB1"

Ignore Patterns

The ignore.columns list supports two forms:

• *.column_name — ignores that column on every table
• table_name.column_name — ignores that column on a specific table only

Ignored columns are excluded from row hashing, so changes to those columns will not appear as conflicts or data differences.

Migration Safety Controls

By default all destructive DDL operations are blocked. When schema-migrate or gen-pack encounters a blocked operation it emits a warning and comments out the statement in the generated SQL. Set the relevant flag to true to allow the operation to be generated as executable SQL.

Key	Controls
allow_drop_column	DROP COLUMN / DROP COLUMN equivalent per driver
allow_drop_table	DROP TABLE
allow_drop_index	DROP INDEX
allow_drop_foreign_key	DROP FOREIGN KEY / DROP CONSTRAINT
allow_modify_primary_key	DROP PRIMARY KEY + ADD PRIMARY KEY

Conflict Resolution Strategies

Strategy	Behaviour
ours	Keep the production row value; discard the dev change
theirs	Replace the production row value with the dev value
manual	Flag for interactive resolution via resolve-conflicts

Performance Options

Key	Default	Description
performance.hash_batch_size	10000	Rows per keyset-paginated query. Set to 0 to load the whole table in one query.
performance.max_parallel_tables	1	Maximum tables hashed concurrently. Increase on multi-core hosts with fast storage.

These values can also be overridden at runtime with --batch-size and --parallel flags on the diff and gen-pack commands.