SQLite strives to be flexible regarding the datatype of the content that it stores. For example, if a table column has a type of "INTEGER", then SQLite tries to convert anything inserted into that column into an integer. So an attempt to insert the string '123' results in an integer 123 being inserted. But if the content cannot be losslessly converted into an integer, for example if the input is 'xyz', then the original string is inserted instead. See the Datatypes In SQLite document for additional information.
Some developers appreciate the freedom that SQLite's flexible typing rules provide and use that freedom to advantage. But other developers are aghast at SQLite's flagrant rule-breaking and prefer the traditional rigid type system found in all other SQL database engines and in the SQL standard. For this latter group, SQLite supports a strict typing mode, as of version 3.37.0 (pending), that is enabled separately for each table.
In a CREATE TABLE statement, if the "STRICT" table-option keyword is added to the end, after the closing ")", then strict typing rules apply to that table. This means:
Every column definition must specify a datatype for that column. The freedom to specify a column without a datatype is removed.
The datatype must be one of following:
No other datatype names are allowed, though new types might be added in future releases of SQLite.
Content inserted into the column with a datatype other than ANY must be either a NULL (assuming there is no NOT NULL constraint on the column) or the type specified. SQLite attempts to coerce the data into the appropriate type using the usual affinity rules, as PostgreSQL, MySQL, SQL Server, and Oracle all do. If the value cannot be losslessly converted in the specified datatype, then an SQLITE_CONSTRAINT_DATATYPE error is raised.
Columns with datatype ANY can accept any kind of data (except they will reject NULL values if they have a NOT NULL constraint, of course). No type coercion occurs for a column of type ANY in a STRICT table.
Columns that are part of the PRIMARY KEY must be NOT NULL.
The ability to host any type of data in a single column has proven to be remarkably useful over the years. In order to continue supporting this ability, even in STRICT tables, the new ANY datatype name is introduced. When the datatype of a column is "ANY", that means that any kind of data - integers, floating point values, strings, or binary blobs, can be inserted into that table and its value and datatype will be preserved exactly as it is inserted. As far as we know, SQLite is the only SQL database engine that supports this advanced capability.
The behavior of ANY is slightly different in a STRICT table versus an ordinary non-strict table. In a STRICT table, a column of type ANY always preserves the data exactly as it is received. For an ordinary non-strict table, a column of type ANY will attempt to convert strings that look like numbers into a numeric value, and if successful will store the numeric value rather than the original string. For example:
CREATE TABLE t1(a ANY) STRICT; INSERT INTO t1 VALUES('000123'); SELECT typeof(a), quote(a) FROM t1; -- result: text '000123'
CREATE TABLE t1(a ANY); INSERT INTO t1 VALUES('000123'); SELECT typeof(a), quote(a) FROM t1; -- result: integer 123
The STRICT keyword at the end of a CREATE TABLE statement is only recognized by SQLite version 3.37.0 (pending) and later. If you try to open a database containing the STRICT keyword in an earlier version of SQLite, it will not recognize the keyword and will report an error (except as noted below). But apart from the extra STRICT keyword, the underlying file format of the database is identical.
Thus, in general, a database file that contains one or more STRICT tables can only be read and written by SQLite version 3.37.0 or later. However, a database created by SQLite 3.37.0 or later can still be read and written by earlier versions of SQLite, going all the way back to version 3.0.0 (2004-06-18) as long as the database does not contain any STRICT tables or other features that were introduced after the older version of SQLite.
Because of a quirk in the SQL language parser, versions of SQLite prior to 3.37.0 can still read and write STRICT tables if they set "PRAGMA writable_schema=ON" immediately after opening the database file, prior to doing anything else that requires knowing the schema. One of the features of PRAGMA writable_schema=ON is that it disables errors in the schema parser. This is intentional, because a big reason for having PRAGMA writable_schema=ON is to facilitate recover of database files with corrupt schemas. So with writable_schema=ON, when the schema parser reaches the STRICT keyword, it says to itself "I don't know what to do with this, but everything up to this point seems like a valid table definition so I'll just use what I have." Hence, the STRICT keyword is effectively ignored. Because nothing else about the file format changes for STRICT tables, everything else will work normally. Of course, rigid type enforcement will not occur because the earlier versions of SQLite do not know how to do that.
The .dump command in the CLI sets PRAGMA writable_schema=ON, because .dump is designed to extract as much content as it can even from a corrupt database file. Hence, if you are using an older version of SQLite and you open a database with STRICT tables in the CLI and issue the ".dump" command before doing anything else, you will be able to read and write to the STRICT tables without rigid type enforcement. This could, potentially, corrupt the database, by allowing incorrect types into STRICT tables. Reopening the database with a newer version of SQLite and running "PRAGMA quick_check" will detect and report all such corruption.
The SQLite parser accepts a comma-separated list of table options after the final close parenthesis in a CREATE TABLE statement. As of this writing (2021-08-23) only two options are recognized:
If there are multiple options, they can be specified in any order. To keep things simple, the current parser accepts duplicate options without complaining, but that might change in future releases, so applications should not rely on it.
*** DRAFT ***