Character data types are strings of characters. Uppercase and lowercase alphabetic characters are accepted literally. The fixed-length character data types are char and c; the variable-length character data types are varchar, long varchar, and text.

The maximum row length depends on the default_page_size setting (a DBMS Server configuration parameter) and can be set to a maximum of 32,767 bytes. For further information on page and row size configuration, see the Database Administrator Guide.

The maximum length of a character column is limited by the maximum row width configured but cannot exceed 32,000 bytes. Long varchar columns are an exception: the maximum length of these columns is 2 GB.

Fixed-length char strings can contain any printing or non-printing character, and the null character ('\0'). In uncompressed tables, char strings are padded with blanks to the declared length. (If the column is nullable, char columns require an additional byte of storage.) For example, if ABC is entered into a char(5) column, five bytes are stored, as follows:

Leading and embedded blanks are significant when comparing char strings. For example, the following char strings are considered different:

When selecting char strings using the underscore (_) wildcard character of the LIKE predicate, include any trailing blanks to be matched. For example, to select the following char string:

the wildcard specification must also contain trailing blanks:

Length is not significant when comparing char strings; the shorter string is (logically) padded to the length of the longer. For example, the following char strings are considered equal:

The c data type is fixed-length and accepts only printing characters. Non-printing characters, such as control characters, are converted into blanks.

Blanks are ignored when c strings are compared, including when compared against other character data types. For example, this c string:

is considered equal to:

Varchar strings are variable-length strings, stored as a 2-byte (smallint) length specifier followed by data. In uncompressed tables, varchar columns occupy their declared length. For example, if ABC is entered into a varchar(5) column, the stored result is:

where:

If the column is nullable, varchar columns require an additional byte of storage.

In compressed tables, varchar columns are stripped of trailing data. For example, if “ABC” is entered into a varchar(5) column in a compressed table, the stored result is:

The varchar data type can contain any character, including non-printing characters and the ASCII null character ('\0').

Except when comparing with C data, blanks are significant in the varchar data type. For example, the following two varchar strings are not considered equal:

and

If the strings being compared are unequal in length, the shorter string is padded with trailing blanks until it equals the length of the longer string.

For example, consider the following two strings:

where:

'\001' represents one ASCII character (ControlA)

and

If they are compared as varchar data types, then

because the blank character added to 'abcd' to make the strings the same length has a higher value than ControlA ('\040' is greater than '\001').

Text strings are variable-length strings. All ASCII characters except the null character (\0) are allowed within text strings. Null characters are converted to blanks.

Blanks are only ignored when text strings are compared with C data. Unlike varchar, if the strings are unequal in length, blanks are not added to the shorter string. For example, assume that the following text strings are being compared:

and

The string 'abcd  ' is considered greater than the string 'abcd' because it is longer.

The long varchar data type has the same characteristics as the varchar data type, but can accommodate strings up to 2 GB in length.

If a length is declared for this type of column, it represents the size of the inline space to store the initial n bytes of data for the large object. For more information, see Storing BLOBs Inline the Database Administrator Guide.

In embedded SQL, data handlers can be created, which are routines to read and write the data for columns that contain large objects. For more information on data handlers, see Data Handlers for Large Objects on page 293 and the Embedded SQL Companion Guide.

The following restrictions apply to long varchar columns: