scanf(3) Library Functions Manual scanf(3)
NAME
scanf, fscanf, sscanf - Converts formatted input
LIBRARY
Standard C Library (libc.so, libc.a)
SYNOPSIS
#include <stdio.h>
int scanf(
const char *format
[,pointer]...);
int fscanf(
FILE *stream,
const char *format
[,pointer]...);
int sscanf(
const char *string,
const char *format
[,pointer]...);
If the pointer parameter identifies an object of type wchar_t (see the Description section), source files should include either
<sys/types.h> or <stddef.h> before <stdio.h> to maintain portability across all systems that conform to current versions of ANSI, ISO, or
X/Open standards.
STANDARDS
Interfaces documented on this reference page conform to industry standards as follows:
fscanf(), scanf(), sscanf(): ISO C, XPG4, XPG4-UNIX
Refer to the standards(5) reference page for more information about industry standards and associated tags.
PARAMETERS
Specifies the format conversion. Specifies the input stream. Specifies input to be read. Points to the location to store the interpreted
data.
DESCRIPTION
The scanf(), fscanf(), and sscanf() functions read character data, interpret it according to a format, and store the converted results into
specified memory locations. The format parameter contains conversion specifications used to interpret the input. The pointer parameters
specify where to store the interpreted data.
The functions read their input from the following sources: Reads from standard input (stdin). Reads from the stream parameter. Reads from
the character string specified by the string parameter.
If the length of an input item is zero, these functions return an error. This error indicates a matching failure unless end-of-file, an
encoding error, or a read error prevented input from a stream, in which case the error indicates input failure.
If there are insufficient arguments for format, the function's behavior is undefined. If format is exhausted while arguments remain, the
excess arguments are evaluated as always but are otherwise ignored.
These functions truncate leading zeroes (before a decimal point, if any). If a string value exceeds its formatted type destination limit
(as defined by the machine architecture), the return value will be the corresponding MIN or MAX value for the type, as appropriate, and an
errno ERANGE value will be set.
The format parameter can contain the following items: A conversion specification that directs the conversion of the next input field. Con-
version specifications start with a % (percent sign). Any white-space character (as determined by the isspace() function) that matches 0
(zero) or more white-space characters in the input stream. Any character except % (percent sign) or a white-space character that must
match the next character in the input stream.
The input stream is broken into fields based on the following: White space
All conversion specifications except %c, %C, and %[ ignore leading white space and consider the first trailing white-space character
as a field delimiter. Invalid character
If the input stream contains a character that is not allowed, this invalid character delimits the field and is considered to be the
first character of the next field. Maximum width
If the conversion specification includes a maximum width and the field is not terminated by white space or an invalid character, the
field is terminated when that character position is reached in the input stream.
Conversion Specifications
Each conversion specification in the format parameter has the following syntax: The character % (percent sign).
The scanf() functions can handle a format string that enables the system to process elements of the pointer list in variable order.
In such a case, the normal conversion character % (percent sign) is replaced by %digit$, where digit is a decimal number in the
range from 1 to NL_ARGMAX. Conversion is then applied to the specified pointer, not to the next unused pointer. This feature pro-
vides for the definition of format strings in an order appropriate to specific languages. If the variable ordering feature is used,
it must be specified for all conversions except for conversion specifications that do not have corresponding pointers (conversion
specifications with the * (asterisk) assignment suppression and %% conversion specifications). If more than one conversion specifi-
cation specifies the same digit, the results of the function are undefined. The optional assignment suppression character * (aster-
isk). An optional decimal digit string that specifies the maximum field width. An optional h or l indicating the size of the
receiving variable for some conversion specifiers, as follows: An h followed by a d, i, o, u, or x conversion specifier indicates
that the receiving variable will be treated as a short int or unsigned short int. An l followed by a d, i, o, u, or x conversion
specifier indicates that the receiving variable will be treated as a long int or unsigned long int. An l followed by an e, f, or g
indicates that the receiving variable will be treated as a double instead of a float. An L followed by an e, f, or g indicates that
the receiving variable will be treated as a long double instead of a float. An l followed by a c, s, or [scanset] indicates that
the receiving variable will be treated as wchar_t instead of char. A conversion code character that specifies the type of conver-
sion to be applied: Accepts a single % (percent sign) input at this point; no assignment is done. Accepts an optionally signed dec-
imal integer, whose format is the same as expected for the subject sequence of strtol() with the value 10 for the base argument.
The pointer parameter should be an integer pointer. Accepts an optionally signed decimal integer, whose format is the same as
expected for the subject sequence of strtol() with the value 0 for the base argument. The pointer parameter should be an integer
pointer. Accepts an unsigned decimal integer; the pointer parameter should be an unsigned integer pointer. Accepts an octal inte-
ger; the pointer parameter should be an integer pointer. Accepts a hexadecimal integer; the pointer parameter should be an integer
pointer. Accepts a floating-point number. The next field is converted accordingly and stored through the corresponding parameter,
which should be a pointer to a float. The input format for floating-point numbers is a string of digits, with the following
optional characteristics: It can be a signed value. It can be an exponential value, containing a decimal point followed by an expo-
nent field, which consists of an E or an e followed by an optionally signed integer. It can be one of the special values INF, NaNQ,
or NaNS. This value is translated into the ANSI/IEEE value for infinity, quiet NaN, or signaling NaN, respectively. Matches an
unsigned hexadecimal long integer, the same as the %p conversion of the printf() function. The corresponding argument should be a
pointer to a pointer to void. No input is consumed. The corresponding argument is a pointer to an integer into which is written the
number of characters read from the input stream by this function. The assignment count returned at the completion of this function
is not incremented. Accepts a string of bytes that are not white-space characters. [ISO C] (When the current locale supports
shift-state encoding, skipping white-space characters may result in redundant shift sequences.)
If no l qualifier is present, the pointer parameter should point to an array of characters that is large enough to accept the con-
verted sequence of characters, along with the terminating null byte automatically appended by the function. When interpreting the
input string, the function considers a white-space character as the delimiter of each input field and generates a string of char
values as output. If a field width is given, the function assumes that pointer refers to a single-byte character array, and only
the specified number of char values is read from the input string.
[ISO C] If an l qualifier is present, the input is treated as a sequence of multibyte characters that begins in the initial shift
state. The function converts each multibyte character to a wide-character as if by a call to the mbrtowc() function, with the con-
version state described by an mbstate_t object initialized to zero before the first multibyte character is converted. The corre-
sponding pointer should point to a wchar_t array that is large enough to accept the converted sequence of wide-characters, plus the
terminating null wide-character that is automatically added by the function. Accepts a string of multibyte characters and converts
them as if by a call to the mbstowcs() function. The pointer parameter should be a pointer to an array of wchar_t. The array must
be large enough to accept the string, along with the terminating null wide-character that is automatically added by the function.
The function treats a white-space character as the delimiter of each field in the input string and generates a string of wchar_t as
output. If the S conversion specifier includes a field width, the behavior of the conversion is undefined. Accepts a sequence of
characters, the number of which is specified by the field width (1 if no field width is specified).
If the l qualifier is not present, the corresponding argument should be a character array large enough to accept the converted
sequence. The function does not append a terminating null character to this sequence.
[ISO C] If the l qualifier is present, the corresponding argument is a sequence of multibyte characters that begins in the initial
shift state. The function converts each multibyte character as if by a call to the mbrtowc() function, with the conversion state
described by an mbstate_t object initialized to zero before conversion of the first multibyte character. The corresponding argument
should be a pointer to the first element of a wchar_t array that is large enough to accept the resulting sequence of wide-charac-
ters. The function does not append a terminating null wide-character to this sequence.
The c directive suppresses the normal skip over white space; therefore, use %1s instead of %1c to read the next nonwhite-space char-
acter. Accepts a single character or a series of characters and converts to wchar_t type. If there is no field width or a field
width of 1 in the conversion specification, one character is accepted and the pointer parameter should be a wchar_t pointer. If
there is a field width greater than 1, the indicated number of characters are accepted and the pointer parameter should be an array
of wchar_t. The normal skip over white space is suppressed. Use %1S rather than %1C to read the next nonwhite-space character.
Accepts as input the characters included in the scanset. The scanset parameter explicitly defines the characters that are accepted
in the string data as those enclosed within [ ] (square brackets).
If the l qualifier is not present, the corresponding pointer parameter should point to an array of char that is large enough to con-
tain both the converted sequence and the terminating null character that is automatically added by the function.
[ISO C] If the l qualifier is present, the input is handled as a sequence of multibyte characters that begins in the initial shift
state. The function converts each multibyte character as if by a call to the mbrtowc() function, with the conversion state described
by an mbstate_t object initialized to zero before conversion of the first multibyte character. The corresponding pointer parameter
should be a pointer to a wchar_t array that is large enough to accept both the converted sequence and the terminating null wide-
character that is automatically added by the function.
The [scanset] directive suppresses the normal skip over leading white space.
A scanset in the form of [^scanset] is an exclusive scanset; that is, the ^ (circumflex) serves as a complement operator and the
characters in scanset are not accepted as input.
Conventions used in the construction of the scanset are as follows: You can represent a range of characters by the construct First-
Last. Thus, you can express [0123456789] as [0-9]. The First parameter must be lexically less than or equal to Last; otherwise, the
- (dash) stands for itself. The dash also stands for itself whenever it is the first or the last character in the scanset. You can
include the ] (right bracket) as an element of the scanset if it is the first character of the scanset. In this case, the right
bracket is not interpreted as the bracket that closes the scanset. If the scanset is an exclusive scanset, the ] character is pre-
ceded by the ^ (circumflex) character to make the ] an element of the scanset.
The conversion specification syntax is summarized by the following synopsis:
%[digit$][*][width][sizecode]convcode
The results from the conversion are placed in *pointer unless you specify assignment suppression with an * (asterisk). Assignment suppres-
sion provides a way to describe an input field that is to be skipped. The input field is a string of nonwhite-space characters. It extends
to the next inappropriate character or until the field width, if specified, is exhausted.
The conversion code indicates how to interpret the input field. The corresponding pointer must usually be of a restricted type. You should
not specify the pointer parameter for a suppressed field.
All *scanf() functions end at the end of the file, at the end of the control string, or when an input character conflicts with the control
string. If the function ends with an input character conflict, the conflicting character is not read from the input stream.
Unless there is a match in the control string, these functions do not read trailing white space (including a newline character).
The success of literal matches and suppressed assignments cannot be directly determined. The *scanf() functions return only the number of
successfully matched and assigned input items.
RESTRICTIONS
Currently, the Tru64 UNIX product does not include locales that use shift-state encoding. Some sections of this reference page refer to the
mb_state object or describe behavior that is dependent on shift-state encoding. This information is included only for your convenience in
developing portable applications that run on multiple platforms, some of which may supply locales that do use shift-state encoding.
RETURN VALUES
The scanf(), fscanf(), and sscanf() functions return the number of successfully matched and assigned input items. This number can be 0
(zero) if there was an early conflict between an input character and the control string. If the input ends before the first conflict or
conversion, the functions return EOF (End-of-File).
ERRORS
The fscanf() function fails if the stream is unbuffered or if the stream's buffer needs to be flushed and the function call causes an
underlying read() or lseek() to be invoked and that operation fails. In addition, the scanf(), fscanf(), and sscanf() functions set errno
to the corresponding value for the following conditions: [Compaq] The O_NONBLOCK flag is set for the underlying stream and the process
would be delayed by the read operation. [Compaq] The file descriptor underlying the stream is not a valid file descriptor or is not open
for reading. The input byte sequence does not form a valid character. [Compaq] The read operation was interrupted by a signal that was
caught and no data was transferred. [Compaq] The call is attempting to read from the process's controlling terminal and either the
process group is orphaned or the process is ignoring or blocking the SIGTTIN signal. [Compaq] Insufficient memory is available for the
operation. [Compaq] One or more of the following errors: The result would exceed the system-defined limits or cause an overflow (value
too large) or an underflow (value too small). The magnitude of x is such that total or partial loss of significance resulted.
RELATED INFORMATION
Functions: atof(3), atoi(3), getc(3), getwc(3), mbstowcs(3), mbtowc(3), printf(3), wprintf(3), wscanf(3)
Standards: standards(5) delim off
scanf(3)