Setting Up a Comparison

Choose the two schemas and which differences you care about.

The setup sheet appears when a comparison starts. It has three parts: the two schemas, the object types to compare, and a long list of differences to ignore.

Choosing the Schemas #

Each side has a connection popup, listing your saved connections, and a database popup. Choosing a connection connects in the background and fills in the databases on that server. The sides are independent, so you can compare two schemas on one server or on two different servers.

Reference is the correct one and is never modified. Target is the one a synchronization would change.

Schema Object Types #

Compare Tables, Views, Functions, Procedures, Triggers, and Events each turn a whole object type on or off. Turn off types you don't care about, such as views when a reporting copy is supposed to have different ones, to keep the results readable.

Ignoring Differences #

Each option marks a kind of difference as not a difference. Use them to turn a report full of noise into one that shows only what you're looking for.

General #

Option	Effect
Ignore case in object names	Treats `Orders` and `orders` as the same object. Needed when comparing across platforms with different default case sensitivity.
Ignore spaces in object names	Ignores whitespace differences in names.
Ignore white space	Ignores formatting differences inside definitions, the usual reason two identical views report as different.
Ignore case	Ignores case differences inside definitions.
Ignore comments	Ignores object and column comments.
Ignore permissions	Leaves privileges out.
Ignore DEFINER and SQL SECURITY clauses	Almost always worth turning on. Definers are usually per-environment users, so otherwise every routine, view, and trigger differs.
Remove occurrences of Reference Database name	Strips the reference schema's name from definitions, so a view that qualifies its tables with the schema name doesn't differ just because the schemas are named differently.
Disable dependency analysis	Turns off the ordering that decides what must be created or dropped before what. Leave it on unless it causes a problem.

Tables #

Option	Effect
Ignore table column order	Compares columns by name, not position. Usually what you want, since reordering columns is expensive and functionally irrelevant.
Associate similar column names	Matches columns with slightly different names, so a rename is reported as a change rather than a drop plus an add.
Ignore column data types, Ignore column default values, Ignore NOT NULL constraints	Narrow the comparison to structure over definition detail.
Ignore table engine, Ignore collations, Ignore auto increment table option, Ignore average row length	Per-environment table options that are rarely meaningful. Auto-increment values differ constantly and never matter.
Ignore partitions	Useful when production is partitioned and development isn't.
Ignore triggers, Ignore triggers present only in Reference, Ignore firing order for triggers	Trigger-specific exclusions.
Ignore check constraints, Ignore constraint names	Ignoring names helps when the server auto-named constraints, so they differ everywhere.
Ignore ENDS clause, Ignore STARTS clause	For events, whose schedule boundaries are usually environment-specific.

Indexes #

Option	Effect
Ignore all indexes	Skips index comparison entirely.
Ignore index names, Ignore column order in indexes	Compares what an index covers, not its name or declared column order.
Ignore primary keys, Ignore foreign keys	Excludes key constraints.
Ignore fulltext indexes, Ignore spatial indexes	Excludes those index types.
Ignore indexes present only in Reference	Doesn't report, or generate SQL to create, indexes the target lacks. Use when the target is deliberately indexed more lightly.

Saving a Setup #

The action popup in the setup sheet has Save Settings as… and Load Settings…, and lists recent settings files. A settings file records both connections, both databases, and every option, so re-running a regular comparison takes two clicks.