Skip to main content

VSQL

Voedger SQL (VSQL) is based on the traditional SQL (Structured Query Language) and used to interact with the Voedger database. Main differences between VSQL and SQL:

  • Alterability: VSQL eliminates the need for the ALTER statement, streamlining schema modifications with automatic, zero downtime updates.
  • Nullability: VSQL tables do not support NULL values, simplifying data handling by avoiding NULL-related operations.
  • Simplified referencing
  • Table inheritance
  • Workspace concept
  • Nested Table concept
  • Singletone concept
  • Projector concept
  • Abstract Table concept
  • Extension Engine concept
  • New data types: Any, raw

Alterability​

VSQL does not use the ALTER TABLE statement. Instead, schema updates are made directly to the TABLE description. During deployment new schemas applied automatically and seamlessly, with zero downtime. Consider an example of schema evolution.

MyTable, version 1:

TABLE MyTable (
    Field1 int NOT NULL,
    Field2 int,
    Field3 int
);

Updated to version 2 by adding Field4:

TABLE MyTable (
    Field1 int,             -- NOT NULL removed
    Field2 int NOT NULL,    -- Added NOT NULL
    Field3 int,
    Field4 int              -- Added Field4
);

Only changes that maintain backward compatibility are permitted. For example, you can't change the type of the existing field, remove a field, or change the order of the fields. Compatibility can be checked at design time with the $ vpm compat command.

Nullability​

Nullability refers to the ability of a database column to accept NULL values.

In VSQL, column values cannot be NULL, eliminating the need for IS NULL, COALESCE(), and other NULL-related functions. Think of Voedger tables as C-structures whose fields have basic types.

However, field can have the NOT NULL attribute, which means that you need to specify a value for this field when inserting a new record.

Please note:

  • Voedger does NOT enforce the NOT NULL attribute when updating records (and, of course, you can not supply the NULL values).
  • Adding a new field with the NOT NULL attribute to an existing table means querying this field in existing records will return a default value: 0 for integer fields, "" (empty string) for string fields, false for boolean fields, etc.