PostgreSQL Query Catalog

Find tables that contain multiple columns for free-form textual descriptions. Make sure that the names of columns are understandable and sufficiently different. Make sure that there are no duplicate columns.

Find inheritance between base tables. Use table inheritance carefully because, for instance, certain constraints are not inherited and must be redefined on child tables.

Find in case of each base table that participates in a table inheritance hierarchy the path to the table from the top-level table. Use table inheritance carefully because, for instance, certain constraints are not inherited and must be redefined on child tables. Also make sure that the identifier of each child table in an inheritance hierarchy is a hyponym of the identifier of its parent table.

Check as to whether there are no unnecessary privileges.

You should follow the principle of least privilege and thus not have in your database tables that usage privileges are granted to the pseudo-role PUBLIC, i.e., to all the database users now and in the future.

Find table, routine, and usage privileges that have been granted to a superuser. Superuser can do anything in the database and thus does not need the privileges. The result is a sign that perhaps the executed GRANT statements were incorrect (wrong username) or the grantee later got superuser status (that it shouldn't have).

Find information about tables that are associated with triggers.

Do not have in a database elements that are not useful. PostgreSQL permits tables with no columns. Such tables can be used to implement Boolean variables (tables TABLE_DEE and TABLE_DUM). On the other hand, such tables might be a result of database evolution, where developers have not noticed that they have dropped all the columns of a table or have not noticed that they have created such a table in the first place.

Find tables with the same name in different schemas. Make sure that this is not a duplication.

Find base table columns with a check constraint that refers to a temporal function (current_timestamp, localtimestamp, current_date, or now) that return type is inconsistent with the data type of the column.

Find indexed textual columns where the indexing does not consider the possibility of pattern-based search. Such columns do not have an index where the used operator class makes the index suitable for use by queries involving pattern matching expressions.

Write correct code. If you pass arguments to a trigger function, then the function should use the arguments. TG_ARGV[]: "Data type array of text; the arguments from the CREATE TRIGGER statement. The index counts from 0. Invalid indexes (less than 0 or greater than or equal to tg_nargs) result in a null value." (PostgreSQL documentation)

Find check constraints of domains where the Boolean expression invokes an operation that does not match with the data type of the domain.

Find surrogate keys where the generator can output the same value more than once. Key values must be unique, i.e., at some point the generator will prevent adding new rows to the table.

Find the TOP 3 longest (identifiers) names of user-defined objects.

"Names in software are 90 percent of what make software readable. You need to take the time to choose them wisely and keep them relevant. Names are too important to treat carelessly. Names should not cause confusion." (Robert C. Martin, Clean Code) Names should be expressive. Find the TOP 3 longest (identifiers) names of user-defined objects by their type. These could be the first candidates of renaming in order to give to database objects better names.

Find the longest referential paths between the tables regardless of the schema that contain the tables.

Perhaps the character maximum length has been omitted accidentally, i.e., one wrote VARCHAR instead of VARCHAR(n) where n is the maximum permitted number of characters in the field value. VARCHAR and TEXT are synonyms.

Find user-defined non-trigger SQL and PL/pgSQL routines where the beginning of the name of the routine indicates a certain action inside the routine (INSERT, UPDATE, or DELETE) but there is no such statement in the routine body.

Find the number and percentage of base tables where all the non-primary key columns are optional