Follow Techotopia on Twitter

On-line Guides
All Guides
eBook Store
iOS / Android
Linux for Beginners
Office Productivity
Linux Installation
Linux Security
Linux Utilities
Linux Virtualization
Linux Kernel
System/Network Admin
Programming
Scripting Languages
Development Tools
Web Development
GUI Toolkits/Desktop
Databases
Mail Systems
openSolaris
Eclipse Documentation
Techotopia.com
Virtuatopia.com
Answertopia.com

How To Guides
Virtualization
General System Admin
Linux Security
Linux Filesystems
Web Servers
Graphics & Desktop
PC Hardware
Windows
Problem Solutions
Privacy Policy

  




 

 

1.9.6.2. Constraints on Invalid Data

Before MySQL 5.0.2, MySQL is forgiving of illegal or improper data values and coerces them to legal values for data entry. In MySQL 5.0.2 and up, that remains the default behavior, but you can change the server SQL mode to select more traditional treatment of bad values such that the server rejects them and aborts the statement in which they occur. Section 5.2.5, “The Server SQL Mode”.

This section describes the default (forgiving) behavior of MySQL, as well as the newer strict SQL mode and how it differs.

If you are not using strict mode, then whenever you insert an “incorrect” value into a column, such as a NULL into a NOT NULL column or a too-large numeric value into a numeric column, MySQL sets the column to the “best possible value” instead of producing an error: The following rules describe in more detail how this works:

  • If you try to store an out of range value into a numeric column, MySQL Server instead stores zero, the smallest possible value, or the largest possible value, whichever is closest to the invalid value. column.

  • For strings, MySQL stores either the empty string or as much of the string as can be stored in the column.

  • If you try to store a string that doesn't start with a number into a numeric column, MySQL Server stores 0.

  • Invalid values for ENUM and SET columns ae handled as described in Section 1.9.6.3, “ENUM and SET Constraints”.

  • MySQL allows you to store certain incorrect date values into DATE and DATETIME columns (such as '2000-02-31' or '2000-02-00'). The idea is that it's not the job of the SQL server to validate dates. If MySQL can store a date value and retrieve exactly the same value, MySQL stores it as given. If the date is totally wrong (outside the server's ability to store it), the special “zero” date value '0000-00-00' is stored in the column instead.

  • If you try to store NULL into a column that doesn't take NULL values, an error occurs for single-row INSERT statements. For multiple-row INSERT statements or for INSERT INTO ... SELECT statements, MySQL Server stores the implicit default value for the column data type. In general, this is 0 for numeric types, the empty string ('') for string types, and the “zero” value for date and time types. Implicit default values are discussed in Section 11.1.4, “Data Type Default Values”.

  • If an INSERT statement specifies no value for a column, MySQL inserts its default value if the column definition includes an explicit DEFAULT clause. If the definition has no such DEFAULT clause, MySQL inserts the implicit default value for the column data type.

The reason for using the preceding rules in non-strict mode is that we can't check these conditions until the statement has begun executing. We can't just roll back if we encounter a problem after updating a few rows, because the storage engine may not support rollback. The option of terminating the statement is not that good; in this case, the update would be “half done,” which is probably the worst possible scenario. In this case, it's better to “do the best you can” and then continue as if nothing happened.

In MySQL 5.0.2 and up, you can select stricter treatment of input values by using the STRICT_TRANS_TABLES or STRICT_ALL_TABLES SQL modes:

SET sql_mode = 'STRICT_TRANS_TABLES';
SET sql_mode = 'STRICT_ALL_TABLES';

STRICT_TRANS_TABLES enables strict mode for transactional storage engines, and also to some extent for non-transactional engines. It works like this:

  • For transactional storage engines, bad data values occurring anywhere in a statement cause the statement to abort and roll back.

  • For non-transactional storage engines, a statement aborts if the error occurs in the first row to be inserted or updated. (When the error occurs in the first row, the statement can be aborted to leave the table unchanged, just as for a transactional table.) Errors in rows after the first do not abort the statement, because the table has already been changed by the first row. Instead, bad data values are adjusted and result in warnings rather than errors. In other words, with STRICT_TRANS_TABLES, a wrong value causes MySQL to roll back all updates done so far, if that can be done without changing the table. But once the table has been changed, further errors result in adjustments and warnings.

For even stricter checking, enable STRICT_ALL_TABLES. This is the same as STRICT_TRANS_TABLES except that for non-transactional storage engines, errors abort the statement even for bad data in rows following the first row. This means that if an error occurs partway through a multiple-row insert or update for a non-transactional table, a partial update results. Earlier rows are inserted or updated, but those from the point of the error on are not. To avoid this for non-transactional tables, either use single-row statements or else use STRICT_TRANS_TABLES if conversion warnings rather than errors are acceptable. To avoid problems in the first place, do not use MySQL to check column content. It is safest (and often faster) to let the application ensure that it passes only legal values to the database.

With either of the strict mode options, you can cause errors to be treated as warnings by using INSERT IGNORE or UPDATE IGNORE rather than INSERT or UPDATE without IGNORE.


 
 
  Published under the terms of the GNU General Public License Design by Interspire