Scilab Website | Contribute with GitLab | Mailing list archives | ATOMS toolboxes
Scilab Online Help
6.0.1 - English

Change language to:
Français - 日本語 - Português - Русский

Please note that the recommended version of Scilab is 2024.0.0. This page might be outdated.
See the recommended documentation of this function

Scilab Help >> Elementary Functions > Floating point > format


set the default output format of decimal numbers


format(mode, width)
mw = format()



formatting mode:

  • 1 or "v": variable adaptative formatting. Scilab chooses either the direct representation (e.g. -143.8757) or the exponential/engineering/scientific notation (e.g. -1.438E+02), in order to cope with each processed value as well as with the required width, with a maximal number of output digits.
  • 0 or "e": The exponential /engineering notation is forced and used for all numbers.


width of numbers = number of output characters used, all included: sign of the mantissa, its digits, decimal separator, exponent symbol, sign and digits of the exponent. Default width = 10. Minimal width in "e" mode = 8.


a vector to set new format

wm(1) is the width

wm(2) is the formatting mode: 0 for 'e' and 1 for 'v'


vector returning the current formatting parameters:

  • mw(1) is the formatting mode: 0 for 'e' and 1 for 'v'
  • mw(2) is the width: Number of characters used to output each number.


For complex numbers, each real and imaginary decimal part is output according to format(..). The total width of a complex number is therefore roughly twice greater (+ 1 character for "i" symbol).
Encoded integers are never output according to the "e"xponential mode, even when this formatting is the default mode set for decimal numbers. For them, the decimal separator "." is skipped. This allows to distinguish them from decimal integers that have a dot.
For wide format(), digits that may be displayed beyond the relative accuracy %eps are set to 0.
The default decimal format set with format(..) is a global parameter: Wherever you set it, it then applies everywhere in Scilab and Xcos.

In the old Scilab versions, in "variable format" mode, vector entries which are less than %eps times the maximum absolute value of the entries were displayed as "0". It is no more the case, the clean function can be used to set negligible entries to zeros.


// format() is not taken into account for encoded integers
// -------------------------------------------------------

// format() impacts console's output, disp(), string(), sci2exp(),...
// --------------------------------------------------------------
a = %pi; p = %pi + %z - %e*%z^2;
a, p
string(a), string(p)
[sci2exp(a) sci2exp(p)]

a, p
string(a), string(p)
[sci2exp(a) sci2exp(p)]
format("v",10); // reset

// format() set the width, not a fixed number of digits:
// ----------------------------------------------------
%pi*1e-217, s = string(%pi*1e-217)
length(s) // the missing char is the implicit "+" sign

// This is useful to get aligned and fully informative display of matrices
[ -1.234 %pi*1e-10 %nan ; %e*1e137 -%inf 54312]

// Since Scilab 6.0, unrelevant digits displayed with wide format() are set to 0:

// format() has a global impact
// ----------------------------
function myfun()
format("v", 10)
// it remains everywhere, wherever it has been set:
format("v",10) // reset

See also

  • numbers properties — determine floating-point parameters
  • string — conversion to string
  • sci2exp — converts an expression to a string
  • disp — displays variables
  • print — prints variables in a file
  • clean — cleans matrices (round to zero small entries)
  • C-like format — mprintf, msprintf, mfprintf conversion specifications
  • mprintf — converts, formats, and writes data to the main scilab window
  • msprintf — converts, formats, and writes data in a string
  • mfprintf — converts, formats, and writes data to a file


6.0 For wide format(), digits that may be displayed beyond the relative accuracy %eps are now set to 0.
Report an issue
<< floor Floating point frexp >>

Copyright (c) 2022-2023 (Dassault Systèmes)
Copyright (c) 2017-2022 (ESI Group)
Copyright (c) 2011-2017 (Scilab Enterprises)
Copyright (c) 1989-2012 (INRIA)
Copyright (c) 1989-2007 (ENPC)
with contributors
Last updated:
Mon Feb 12 19:26:47 CET 2018