PostgreSQL Query Catalog

Find cases where the list of valid data values in the column is specified in the column definition (in addition to specifying the type of the column) by using, for instance, check constraints or enumerated types. The check constraint is either associated directly with a table or is associated with a domain.

Find base tables that implement a highly generic database design (EAV design - Entiry-Attribute-Value design), according to which attribute values are recorded in a generic table that contains attribute-value pairs.

Find base base tables have the simple primary key that contains the column with the (case insensitive) name id and an integer type. In addition, the primary key values are generated automatically by the system by using a sequence generator.

Find base base tables have a simple unique index (not associated with a constraint) that contains the column with the (case insensitive) name id and an integer type. In addition, the key values are generated automatically by the system by using a sequence generator.

Find cases where the same column of a base table T is used to record references to multiple base tables. In addition, one has to add additional column to T for holding metadata about the parent table, referenced by the current row.

Find base table columns that have FLOAT, REAL, or DOUBLE PRECISION type. "The data types real and double precision are inexact, variable-precision numeric types. On all currently supported platforms, these types are implementations of IEEE Standard 754 for Binary Floating-Point Arithmetic (single and double precision, respectively), to the extent that the underlying processor, operating system, and compiler support it." (PostgreSQL documentation) Do not use the approximate numeric types FLOAT, REAL, and DOUBLE PRECISION in order to present fractional numeric data. Due to the use of the IEEE 754 standard the results of calculations with the values, which have one of these types, can be inexact because out of necessity some numbers must be rounded to a value, which is very close. "Comparing two floating-point values for equality might not always work as expected." (PostgreSQL documentation)

Find base table columns with the Money data type. Each value of the money type has associated currency sign that depends on server settings. It could be $. Moreover, using the values for arithmetic operations requires casts that makes the code more complicated.

This query identifies superfluous CHECK constraints where a programmatic length check duplicates a declarative, data type-based length limit. For instance, a CHECK constraint like char_length(column) <= 100 on a column already defined as VARCHAR(100) is redundant.

Write code that is simple to understand and not confusing. A double negative is a grammatical construction occurring when two forms of negation are used in the same expression (https://en.wikipedia.org/wiki/Double_negative). Double negatives in Boolean expressions make it more difficult to understand and maintain the code.

Fing regular expression patterns that use [^\S] instead of \s or [^\D] instead of \d or [^\W] instead of \w.

Find default values that return current timestamp with the maximum number of fractional seconds (6).

The same table should not have multiple CHECK constraints with exactly the same Boolean expression. Do remember that the same task can be solved in SQL usually in multiple different ways. Thus, the exact copies are not the only possible duplication.

The same domain should not have multiple CHECK constraints with exactly the same Boolean expression. Do remember that the same task can be solved in SQL usually in multiple different ways. Thus, the exact copies are not the only possible duplication.

Find comments that have been registered with a COMMENT statement and that are associated with more than one object. It would probably mean that a comment is incorrect or missing.

Find base table columns that have both default value determined through a domain and default value that is directly attached to the column. Do not duplicate specifications of default values to avoid confusion and surprises. If column and domain both have a default value, then in case of inserting data the default value that is associated directly with the column is used.

Find domains that have the same properties (base type, character length, not null + check constraints, default value, collation). There should not be multiple domains that have the same properties. Do remember that the same task can be solved in SQL usually in multiple different ways. Therefore, the domains may have syntactically different check constraints that solve the same task. Thus, the exact copies are not the only possible duplication.

Find enumerated types with exactly the same values. There should not be multiple types that have the same values.

Find duplicate foreign key constraints, which involve the same columns and refer to the same set of columns.

Find composite types with the same attributes (regardless of the order of attributes). Make sure that there is no duplication.

Find completely overlapping key (primary key, unique, and exclude where all operators are =) constraints. This is a form of duplication. It leads to the creation of multiple indexes to the same set of columns.