[ Term I/O | Reference Manual | Alphabetic Index ]

printf(+Stream, +Format, ?ArgList)

The arguments in the argument list ArgList are interpreted according to the Format string and the result is printed on the output Stream

Stream: Stream handle or alias (atom)
Format: String or Atom.
ArgList: List or any Term.

Description

Format is an atom or a string which can contain embedded format control sequences. ArgList is a list of arguments that will be interpreted, and possibly printed, under control of these format options. Any part of Format that is not a control sequence is written to the output stream Stream unchanged. As far as the functionality overlaps, the behaviour is similar to the 'C' function printf().

Format control sequences take one of the following forms:

       %[-+ 0][minwidth][.precision]{d|u|o|x|X}
       %[-+ 0][minwidth][.precision]{f|g|e|E}
       %[-+ 0][minwidth.]radix{r|R}
       %[-][minwidth][.precision]{s|a|A}
       %[depth][modifiers]{w|W}
       %[count]{c|i|n|t}
       %{b|k|p|q|%}

Each control sequence starts with a percent character, followed by a number of optional flags/parameters, and ends in one of the conversion control characters (described below in detail).

The meaning of the flags and parameters are:

-: Left-justify the output (if the field width is greater than the length of the printed representation). Default is right-justified.
+: Print a plus sign in front of non-negative numbers. By default, only negative numbers are printed with a sign.
(space): Print a blank space in front of non-negative numbers. If a plus sign is specified as well, the space is ignored.
0: If the width of a right-justified field is greater than the length of the printed representation, fill the space with leading zeros instead of spaces.
minwidth: An optional minimum field width, which causes the printed representation to be padded (left or right) in case it is shorter than the specified minimum. The width is given either as an unsigned decimal integer, or as a '*' character, in which case the number is taken from the next element of ArgList.
precision: An optional precision, determining the number of places after the decimal point (for floating point arguments), the minimum number of digits (for integer arguments), or the maximum number of characters (for atom/string arguments). The precision is given either as an unsigned decimal integer, or as a '*' character, in which case the number is taken from the next element of ArgList.
radix: A number between 2 and 36, giving the number base in which the argument is to be printed. For instance, 2 for binary, 16 for hexadecimal, 8 for octal, or 36 to use all the digits 0-9a-z. The radix is given either as a decimal integer, or as a '*' character, in which case it is taken from the next element of ArgList.
count, depth: An optional count (determining how many times an option will be repeated), or depth (determining the maximum nesting depth for printing terms). It is given either as an unsigned decimal integer, or as a '*' character, in which case the number is taken from the next element of ArgList.

The meaning of the conversion control characters is:

a

The argument has to be an atom (or string) and is printed without any quoting. Optionally, field width, justification, and maximum length can be specified. Identical to the s format.

A

The argument has to be an atom (or string), is converted to upper case, and then printed without quoting. Optionally, field width, justification, and maximum length can be specified.

b

The stream's output buffer is flushed, buffered data is written to the device attached to the stream.

c

The argument has to be an integer character code and is printed as a character. Optionally, a repeat count can be specified.

d

The argument has to be an integer and is printed in signed decimal notation. Optionally, details regarding field width, number of digits, padding and sign representation can be specified. See r for non-decimal notation.

e

The argument has to be a floating-point number and is printed according to the substring A in exponential notation.('e' is used for exponentiation)

E

The argument has to be a floating-point number and is printed according to the substring A in exponential notation. ('E' is used for exponentiation)

f

The argument has to be a floating-point number and is printed according to the substring A in non-exponential form.

g

The argument has to be a floating-point number and is printed according to the substring A in exponential or non-exponential form, whichever gives the best precision in minimum space.('e' is used for exponentiation)

i

The next argument of ArgList is ignored. An optional repeat count can be specified.

k

The argument is passed to display/1. It is a synonym for %O.w.

n

A newline sequences is printed. An optional repeat count can be given. Which newline characters are printed depends on the setting of the stream's end_of_line property. If the stream's flush-property is set to end_of_line, the stream is also flushed.

o

The argument has to be an integer and its least significant 32 bits are printed in unsigned octal notation. Optionally, details regarding field width, number of digits, padding and sign representation can be specified. DEPRECATED, see r.

r

The argument has to be an integer and is printed in signed notation in the given base (radix) between 2 and 36. The digits are represented as the characters 0-9a-z. Optionally, details regarding field width, padding and sign representation can be specified.

R

The argument has to be an integer and is printed in signed notation in the given base (radix) between 2 and 36. The digits are represented as the characters 0-9A-Z. Optionally, details regarding field width, padding and sign representation can be specified.

p

The argument is passed to print/1. It is a synonym for %Pw.

q

The argument is passed to writeq/1. It is a synonym for %QDvw.

s

The argument has to be an string or atom and is printed without any quoting. Optionally, field width, justification, and maximum length can be specified.

t

A tab character is printed. An optional repeat count can be given.

u

The argument has to be an integer and its least significant 32 bits are printed in unsigned decimal notation. Optionally, details regarding field width, number of digits, padding and sign representation specified. DEPRECATED.

w

The argument is by default passed to write/1. However, the %w format recognises a number of modifier characters, placed between the percent sign and w. They give the user full control over the various possibilities of printing Prolog terms. A number immediately after the percent sign determines the depth to which the term is printed, if an asterisk is used instead, the depth is taken from the next argument in ArgList. The default depth is determined by the setting of the (stream-specific or global) print_depth flag. After the optional depth, the following modifiers are recognized:

O: omit operator declarations. All terms are written in the canonical notation without operators.
Q: quote atoms and strings if necessary.
.: write lists in the dot functor notation rather than using the square bracket notation, e.g. .(1, .(2, [])) rather than [1, 2].
G: print the term as a goal, i.e. goal write transformations will be taken into account.
P: call the user-defined predicate portray/1, 2 in the way print/1, 2 does.
D: disregard the depth restriction of the print-depth flag and print the whole term.
U: call portray/1, 2 even on variables. This is to be used in conjunction with the P option. Note that attributed variables are always portrayed.
V: print the full variable name, if available, either in the form Name_Number, e.g. Alpha_132, or Name#Number, if the variable had been given a name via lib(var_name). This is necessary to distinguish different variables with the same name.
v: print only the short variable form, i.e. even when available, the variable name is not printed. This is useful if a term should be written and read back in several times. If neither V nor v is specified, variables are printed only with their name, if it is available. Variable without names are always printed in the v form.
_: print every variable as a simple underscore. Any information about multiple occurrences of a variable is lost with this format. It is mainly useful to produce output that can be compared easily with the output of a different Eclipse session.
I: any term of the form '$VAR'(N), where N is a non-negative integer, is printed as a variable name consisting of a capital letter followed by a number. The capital letter is the ((N mod 26)+1)st letter of the alphabet, and the integer is N//26. If N is an atom, this atom gets printed instead of the term.
K: don't print blank space (around operators, after commas, etc.) unless necessary.
M: print the full contents of all variable attributes. This is necessary if the term is to be written out and read back in.
m: variable attributes are printed using the corresponding print handlers. If neither M nor m is specified, attributed variables are printed as variables, without any attribute.
N: print newline (NL) characters as newlines rather than as an escape sequence, even when they occur in quoted atoms or strings. This only makes sense together with the Q modifier.
T: do not apply any write transformations.
C: print the term as a clause, i.e. clause macros will be taken into account.
F: print a fullstop after the term, separated from the term by an extra space, if necessary.
L: print a newline after the term (or as part of the fullstop sequence, if used together with the F option).

W

Like %w, but the stream's default output options are taken into account, unless overridden by the format options specified here. Note in particular that a default setting may be cancelled by prefixing the format character with a minus sign. E.g. if the stream defaults specify that quotes should be printed (quoted(true)), this can be overridden by a %-QW format string.

x

The argument has to be an integer and its least significant 32 bits are printed in unsigned hexadecimal notation (using letters abcdef). Optionally, details regarding field width, number of digits, padding and sign representation can be specified. DEPRECATED, see r.

X

The argument has to be an integer and its least significant 32 bits are printed in unsigned hexadecimal notation (using letters ABCDEF). Optionally, details regarding field width, number of digits, padding and sign representation can be specified. DEPRECATED, see R.

%

One % is printed.

For backward compatibility, if ArgList contains only one argument, it need not be in a list.

Modes and Determinism

printf(+, +, ?) is det

Modules

This predicate is sensitive to its module context (tool predicate, see @/2).

Exceptions

(4) instantiation fault: Stream is not instantiated.
(5) type error: Stream is not an atom or a stream handle.
(5) type error: Format is not an atom or a string.
(5) type error: ArgList contains argument whose type does not correspond to the control sequence.
(7) string contains unexpected characters: Format is not correct, it contains too many asterisks or a control character is missing or there is a redundant character before the control character.
(8) bad argument list: ArgList has not enough or too many arguments.
(192) illegal stream mode: Stream is not an output stream.
(193) illegal stream specification: Stream is not a stream specification.

Examples