Manual for Package pgfplots
2D/3D Plots in LATeX, Version 1.18.1
http://sourceforge.net/projects/pgfplots

/tikz/mark size={dimension} (initially 2pt) ¶

This TikZ option allows to set marker sizes to dimension. For circular markers, dimension is the radius, for other plot marks it is about half the width and height.

/pgfplots/cube/size x={dimension} (initially \pgfplotmarksize=2pt) ¶

/pgfplots/cube/size y={dimension} (initially \pgfplotmarksize=2pt) ¶

/pgfplots/cube/size z={dimension} (initially \pgfplotmarksize=2pt) ¶

Sets the size for mark=cube separately for every axis.

/tikz/every mark(no value) ¶

This TikZ style can be reconfigured to set marker appearance options like colors or transformations like scaling or rotation. pgfplots appends its cycle list options to this style.



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[y=2cm] \addplot coordinates {(-2,0) (-1,1) (0,0) (1,1) (2,0)}; \end{axis} \end{tikzpicture}



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \tikzset{every mark/.append style={scale=2}} \begin{tikzpicture} \begin{axis}[y=2cm] \addplot coordinates {(-2,0) (-1,1) (0,0) (1,1) (2,0)}; \end{axis} \end{tikzpicture}



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[y=2cm] \addplot+ [ mark=halfcircle*, every mark/.append style={rotate=90}] coordinates {(-2,0) (-1,1) (0,0) (1,1) (2,0)}; \addplot+ [ mark=halfcircle*, every mark/.append style={rotate=180}] coordinates {(-2,-0.1) (-1,0.9) (0,-0.1) (1,0.9) (2,-0.1)}; \end{axis} \end{tikzpicture}

Note that every mark is kind of static in the sense that it is evaluated once only. If you need individually colored marks as part of a scatter plot, you will need to resort to scatter/use mapped color.



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[y=2cm] \addplot+ [ mark=ball, mark size=4pt, scatter,% enable scatter scatter src=rand,% the "color data" % configure individual appearances: scatter/use mapped color= {ball color=mapped color}] coordinates {(-2,0) (-1,1) (0,0) (1,1) (2,0)}; \end{axis} \end{tikzpicture}

/pgfplots/no markers(style, no value) ¶

A key which overrides any mark value set by cycle list of option lists after \addplot.

If this style is provided as argument to a complete axis, it is appended to every axis plot post such that it disables markers even for cycle lists which contain markers.

/tikz/mark repeat={integer} (initially empty) ¶

Allows to draw only each \(n\)th mark where \(n\) is provided as integer.

/tikz/mark phase={integer \(p\)} (initially 1) ¶

This option allows to control which markers are drawn. It is primarily used together with the TikZ option mark repeat=\(r\): it tells TikZ that the first mark to be drawn should be the \(p\)th, followed by the \((p + r)\)th, then the \((p + 2r)\)th, and so on.



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[tiny] \addplot+ [scatter] {sin(deg(x))}; \end{axis} \end{tikzpicture} \begin{tikzpicture} \begin{axis}[tiny] \addplot+ [scatter, mark repeat=3,mark phase=2] {sin(deg(x))}; \end{axis} \end{tikzpicture}

Here, \(p=1\) is the first point (the one with \coordindex\(=0\)).

/tikz/mark indices={index list} (initially empty) ¶

Allows to draw only the marker whose index numbers are in the argument list.

/pgf/mark color={color} (initially empty) ¶

Defines the additional fill color for the halfcircle, halfcircle*, halfdiamond* and halfsquare* markers. An empty value uses white (which is the initial configuration). The value none disables filling for this part.

These markers have two distinct fill colors, one is determined by fill as for any other marker and the other one is mark color.



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[y=2cm] \addplot [ blue, mark color=blue!50!white, mark=halfcircle* ] coordinates { (-2,0) (-1,1) (0,0) (1,1) (2,0) }; \addplot [ red, mark color=red!50!white, mark=halfsquare*, ] coordinates { (-2,-0.1) (-1,0.9) (0,-0.1) (1,0.9) (2,-0.1) }; \end{axis} \end{tikzpicture}

Note that this key requires pgf 2.10 or later.

/tikz/mark options={options} ¶

Resets every mark to {options}.

/pgf/text mark={text} (initially p) ¶

Changes the text shown by mark=text.

With /pgf/text mark=m:

With /pgf/text mark=A:

There is no limitation about the number of characters or whatever. In fact, any TeX material can be inserted as text, including images.

/pgf/text mark style={options for mark=text} ¶

Defines a set of options which control the appearance of mark=text.

If /pgf/text mark as node=false (the default), options is provided as argument to \pgftext – which provides only some basic keys like left, right, top, bottom, base and rotate.

If /pgf/text mark as node=true, options is provided as argument to \node. This means you can provide a very powerful set of options including anchor, scale, fill, draw, rounded corners, etc.

/pgf/text mark as node=true|false (initially false) ¶

Configures how mark=text will be drawn: either as \node or as \pgftext.

The first choice is highly flexible and possibly slow, the second is very fast and usually enough.

\pgfdeclareplotmark{plot mark name}{code} ¶

Defines a new marker named plot mark name. Whenever it is used, code will be invoked. It is supposed to contain (preferable pgf basic level) drawing commands. During code, the coordinate system’s origin denotes the coordinate where the marker shall be placed.

Please refer to the PGF/TikZ manual section “Mark Plot Handler” for more detailed information.

/pgfplots/every axis plot post(style, initially ) ¶

The every axis plot post style can be used to overwrite parts (or all) of the drawing styles which are assigned for plots.



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} % Overwrite any cycle list: \pgfplotsset{ every axis plot post/.append style={ mark=triangle, every mark/.append style={rotate=90}, }, } \begin{tikzpicture} \begin{axis}[y=2cm] \addplot coordinates { (-2,0) (-1,1) (0,0) (1,1) (2,0) }; \end{axis} \end{tikzpicture}

/tikz/solid(style, no value) ¶

/tikz/dotted(style, no value) ¶

/tikz/densely dotted(style, no value) ¶

/tikz/loosely dotted(style, no value) ¶

/tikz/dashed(style, no value) ¶

/tikz/densely dashed(style, no value) ¶

/tikz/loosely dashed(style, no value) ¶

/tikz/dashdotted(style, no value) ¶

/tikz/densely dashdotted(style, no value) ¶

/tikz/loosely dashdotted(style, no value) ¶

/tikz/dashdotdotted(style, no value) ¶

/tikz/densely dashdotdotted(style, no value) ¶

/tikz/loosely dashdotdotted(style, no value) ¶

/tikz/line cap=round|rect|butt (initially butt) ¶

/tikz/line join=round|bevel|miter (initially miter) ¶

/tikz/miter limit=factor (initially 10) ¶

These keys control how lines are joined at edges. Their description is beyond the scope of this manual, so interested readers should consultthe PGF/TikZ manual.

Here is just an example illustrating why it might be of interest to study these parameters:



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} % requires \usetikzlibrary{spy} \begin{tikzpicture}[spy using outlines= {circle, magnification=6, connect spies}] \begin{axis}[no markers,grid=major, every axis plot post/.append style={thick}] \addplot coordinates {(0, 0) (0, 0.9) (1, 0.9) (2, 1) (3, 0.9) (80, 0)}; \addplot+ [line join=round] coordinates {(0, 0) (0, 0.9) (2, 0.9) (3, 1) (4, 0.9) (80, 0)}; \addplot+ [line join=bevel] coordinates {(0, 0) (0, 0.9) (3, 0.9) (4, 1) (5, 0.9) (80, 0)}; \addplot+ [miter limit=5] coordinates {(0, 0) (0, 0.9) (4, 0.9) (5, 1) (6, 0.9) (80, 0)}; \coordinate (spypoint) at (3,1); \coordinate (magnifyglass) at (60,0.7); \end{axis} \spy [blue, size=2.5cm] on (spypoint) in node[fill=white] at (magnifyglass); \end{tikzpicture}

/tikz/font={font name} (initially \normalfont) ¶

Sets the font which is to be used for text in nodes (like tick labels, legends or descriptions).

A font can be any LaTeX argument like \footnotesize or \small\bfseries.³⁵

It may be useful to change fonts only for specific axis descriptions, for example using

copy \pgfplotsset{ tick label style={font=\small}, label style={font=\small}, legend style={font=\footnotesize}, }

See also the predefined styles normalsize, small and footnotesize in Section 4.10.2.

/tikz/line width={dimension} (initially 0.4pt) ¶

Sets the line width. Please note that line widths for tick lines and grid lines are predefined, so it may be necessary to override the styles every tick and every axis grid.

The line width key is changed quite often in TikZ. You should use

copy \pgfplotsset{every axis/.append style={line width=1pt}}

or

copy \pgfplotsset{every axis/.append style={thick}}

to change the overall line width. To also adjust ticks and grid lines, one can use

copy \pgfplotsset{every axis/.append style={ line width=1pt, tick style={line width=0.6pt}}}

or styles like

copy \pgfplotsset{every axis/.append style={ thick, tick style={semithick}}}

The ‘every axis plot’ style can be used to change line widths for plots only.

/tikz/thin(no value) ¶

/tikz/ultra thin(no value) ¶

/tikz/very thin(no value) ¶

/tikz/semithick(no value) ¶

/tikz/thick(no value) ¶

/tikz/very thick(no value) ¶

/tikz/ultra thick(no value) ¶

These TikZ styles provide different predefined line widths.

/tikz/color={a color} ¶

/tikz/draw={stroke color} ¶

/tikz/fill={fill color} ¶

These keys are (generally) used to set colors. Use color to set the color for both drawing and filling. Instead of color={color name} you can simply write color name. The draw and fill keys only set colors for stroking and filling, respectively.

Use draw=none to disable drawing and fill=none to disable filling.³⁶

Since these keys belong to TikZ, the complete documentation can be found in the PGF/TikZ manual, Section “Specifying a Color”.

/pgfplots/colormap name={color map name} (initially hot) ¶

Changes the current color map to the already defined map named color map name. The predefined color maps are

.
viridis
hot

The definition can be found in the documentation for colormap/hot and colormap/viridis, respectively. These, and further color maps, are described below.

Color maps can be used, for example, in scatter plots and surface plots.

You can use colormap to create new color maps (see below).

/pgfplots/colormap={name}{color specification} ¶

Defines a new colormap named name according to color specification and activates it using colormap name={name}.

A simple color specification is just a sequence of color definitions of type

color type=(color value)

separated by either white spaces or semicolon or comma:

copy \pgfplotsset{colormap={CM}{rgb=(0,0,1) color=(black) rgb255=(238,140,238)}}

Here, the three input colors form the left end, middle point, and right end of the interval, respectively. A couple of color types are available, the color value depends on the actual color type which is shown below in all detail. Most colormap definitions use the simple form and merely list suitable color definitions.

A more advanced color specification is one which defines both colors and positions in order to define the place of each color. In this case, pgfplots offers the syntax

color type(offset)=(color value):

copy \pgfplotsset{colormap={CM}{rgb(-500)=(0,0,1) color(0)=(black) rgb255(1500)=(238,140,238)}}

This syntax allows to distribute colors over the interval using nonuniform distances.

Compatibility note:

pgfplots up to and including version \(1.13\) offered just rudimentary support for nonuniform color maps. You need to write compat=1.14 or higher in order to make use of nonuniform color maps.

The positions can be arbitrary numbers (or dimensions),³⁷ but each new color must come with a larger position than its preceding one. The position can be omitted in which case it will be deduced from the context: if the first two colors have no position, the first will receive position \(0\) and the second will receive position “\(1\text {cm}\)”. All following ones receive the last encountered mesh width. Note that nonuniform positions make a real difference in conjunction with colormap access=piecewise const.

The precise input format is described in the following section.

4.7.6.1Colormap Input Format Reference¶

Each entry in color specification has the form

color model(position)=(arguments) or

special mode(position)=(argument of colormap name)

where the most common form is to specify a color using a color model like rgb=(0,0.5,1). The special modes are discussed later in Section 4.7.6.4 on page (page for section 4.7.6.4); they are useful to access colors of existing colormaps. The position argument is optional and defaults to \(0\) for the first color, \(1\text {cm}\) for the second, and an automatically deduced mesh grid for all following ones. The number range of position is arbitrary. Note that pgfplots merely remembers the relative distances of the positions, not their absolute values. Consequently, a color map with positions \(0,1,2\) is equivalent to one with \(0,10,20\) or \(-10,0,10\) or \(-1100,-1000,-900\). pgfplots maps the input positions to the range \([0,1000]\) internally and works with these numbers. Each new color must have a position which is larger than the preceding one.

Available choices for color model are

rgb which expects arguments of the form (red,green,blue) where each component is in the interval \([0,1]\),

rgb255 which is similar to rgb except that each component is expected in the interval [0,255],

gray in which case arguments is a single number in the interval \([0,1]\),

color in which case arguments contains a predefined (named) color like ‘red’ or a color expression like ‘red!50’,

cmyk which expects arguments of the form (cyan,magenta,yellow,black) where each component is in the interval \([0,1]\),

cmyk255 which is the same as cmyk but expects components in the interval \([0,255]\),

cmy which expects arguments of the form (cyan,magenta,yellow) where each component is in the interval \([0,1]\),

hsb which expects arguments of the form (hue,saturation,brightness) where each component is in the interval \([0,1]\),

Hsb which is the same as hsb except that hue is accepted in the interval \([0,360]\) (degree),

HTML which is similar to rgb255 except that each component is expected to be a hex number between 00 and FF,

wave which expects a single wave length as numeric argument in the range \([363,814]\).

The choice of positions influences the processing time: a uniform distance between the positions allows more efficient lookup than nonuniform distances. If there is no visual difference and it does not hurt with respect to the number of data points, prefer color maps with uniform distances over nonuniform maps. Note that nonuniform maps make a huge difference in conjunction with colormap access=piecewise constant. pgfplots provides a simple way to map a nonuniform color definition to a uniform one: write the target mesh width as first item in the specification. pgfplots will perform this interpolation automatically, provided all encountered positions can be mapped to the target grid:

In this last example, the mesh width has been provided explicitly and pgfplots interpolates the missing grid points on its own. It is an error if the provided positions are no multiple of the mesh width.

4.7.6.2The Colorspace of a Colormap¶

Attention: this section is essentially superfluous if you have configured the xcolor package to override color spaces globally (for example by means of \usepackage[cmyk]{xcolor} before loading pgfplots), see the end of this subsection.

Even though a colormap accepts lots of color spaces on input (in fact, it accepts most or all that xcolor provides), the output color of a colorspace has strict limitations. The output colorspace is the one in which pgfplots interpolates between two other colors. To this end, it transforms input colors to the output color space. The output colorspace is also referred to as “the colorspace of a colormap”.

There are three supported color spaces for a colormap: the GRAY, RGB, and CMYK color spaces. Each access into a colormap requires linear interpolation which is performed in its color space. Color spaces make a difference: colors in different color spaces may be represented differently, depending on the output device. Many printers use CMYK for color printing, so providing CMYK colors might improve the printing quality on a color printer. The RGB color space is often used for display devices. The predefined colormaps in pgfplots all use RGB.

Whenever a new colormap is created, pgfplots determines an associated color space. Then, each color in this specific colormap will be represented in its associated color space (converting colors automatically if necessary). Furthermore, every access into the colormap will be performed in its associated color space and every returned mapped color will be represented with respect to this color space. Furthermore, every shading generated by shader=interp will be represented with respect to the colormap’s associated color space.

The color space is chosen as follows: in case colormap default colorspace=auto (the initial configuration), the color space depends on the first encountered color in color specification. For rgb or gray or color, the associated color space will be RGB (as it was in all earlier versions of pgfplots). For cmyk, the associated color space will be CMYK. If colormap default colorspace is either gray, rgb or cmyk, this specific color space is used and every color is converted automatically.

4.7.6.3Predefined Colormaps¶

Available color maps are shown below.

/pgfplots/colormap/viridis(style, no value) ¶

A style which installs the colormap “viridis” which has been defined by Stèfan van der Walt and Nathaniel Smith for Matplotlib. It is designated to be the default colormap for Matplotlib starting with version 2.0 and is released under the CC0³⁸.

The choice viridis is a downsampled copy included in pgfplots.

copy \pgfplotsset{ colormap name=viridis, }

This colormap has considerably better properties compared to other choices:

• • its color distribution is perceptually uniform (compare the definition in the link below),

• • it is suitable for moderate forms of color blindness,

• • it is still good when printed in black and white.

Details about these properties can be found in http://bids.github.io/colormap.

Please use the choice colormap name=viridis as this makes uses of the predefined colormap whereas colormap/viridis will redefine it.

There is also a high resolution copy of viridis which is called colormap/viridis high res in the colormaps library. It resembles the original resolution of the authors, but it is visually almost identically to viridis and requires less resources in TeX.

/pgfplots/colormap/hot(style, no value) ¶

A style which installs the colormap

copy \pgfplotsset{ colormap={hot}{color(0cm)=(blue); color(1cm)=(yellow); color(2cm)=(orange); color(3cm)=(red)} }

This is a pre-configured color map.

/pgfplots/colormap/hot2(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ /pgfplots/colormap={hot2}{[1cm]rgb255(0cm)=(0,0,0) rgb255(3cm)=(255,0,0) rgb255(6cm)=(255,255,0) rgb255(8cm)=(255,255,255)} }

Note that this particular choice ships directly with pgfplots, you do not need to load the colormaps library for this value.

This colormap is similar to one shipped with Matlab® under a similar name.

/pgfplots/colormap/jet(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ /pgfplots/colormap={jet}{rgb255(0cm)=(0,0,128) rgb255(1cm)=(0,0,255) rgb255(3cm)=(0,255,255) rgb255(5cm)=(255,255,0) rgb255(7cm)=(255,0,0) rgb255(8cm)=(128,0,0)} }

This colormap is similar to one shipped with Matlab® under a similar name.

/pgfplots/colormap/blackwhite(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ colormap={blackwhite}{gray(0cm)=(0); gray(1cm)=(1)} }

/pgfplots/colormap/bluered(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ colormap={bluered}{ rgb255(0cm)=(0,0,180); rgb255(1cm)=(0,255,255); rgb255(2cm)=(100,255,0); rgb255(3cm)=(255,255,0); rgb255(4cm)=(255,0,0); rgb255(5cm)=(128,0,0)} }



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[colormap/bluered] \addplot+ [scatter, scatter src=x,samples=50] {sin(deg(x))}; \end{axis} \end{tikzpicture}

Remark:

The style bluered (re)defines the color map and activates it. TeX will be slightly faster if you call \pgfplotsset{colormap/bluered} in the preamble (to create the color map once) and use colormap name=bluered whenever you need it. This remark holds for every color map style which follows. But you can simply ignore this remark.

/pgfplots/colormap/cool(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ colormap={cool}{rgb255(0cm)=(255,255,255); rgb255(1cm)=(0,128,255); rgb255(2cm)=(255,0,255)} }

/pgfplots/colormap/greenyellow(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ colormap={greenyellow}{rgb255(0cm)=(0,128,0); rgb255(1cm)=(255,255,0)} }

/pgfplots/colormap/redyellow(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ colormap={redyellow}{rgb255(0cm)=(255,0,0); rgb255(1cm)=(255,255,0)} }

/pgfplots/colormap/violet(style, no value) ¶

A style which is equivalent to

copy \pgfplotsset{ colormap={violet}{rgb255=(25,25,122) color=(white) rgb255=(238,140,238)} }

\pgfplotscolormaptoshadingspec{colormap name}{right end size}{\macro} ¶

A command which converts a colormap into a pgf shading’s color specification. It can be used in commands like \pgfdeclare...shading (see the PGF/TikZ manual for details).

The first argument is the name of a (defined) colormap, the second the rightmost dimension of the specification. The result will be stored in \macro.

copy % convert `hot' -> \result \pgfplotscolormaptoshadingspec{hot}{8cm}\result % define and use a shading in pgf: \def\tempb{\pgfdeclarehorizontalshading{tempshading}{1cm}}% % where `\result' is inserted as last argument: \expandafter\tempb\expandafter{\result}% \pgfuseshading{tempshading}%

The usage of the result \macro is a little bit low-level.

Attention:

pgf shadings are always represented with respect to the RGB color space. Consequently, even CMYK colormap names will result in an RGB shading specification when using this method.³⁹

\pgfplotscolorbardrawstandalone[options] ¶

A command which draws a tikzpicture and a colorbar using the current colorbar settings inside of it. Its purpose is to simplify the documentation.

Since this colorbar is a “standalone” picture, it defines the following options

copy point meta min=0, point meta max=1000, parent axis width/.initial=6cm, parent axis height/.initial=6cm,

before it evaluates options and draws the colorbar.

\pgfplotscolormaptodatafile[options]{colormap name}{output file} ¶

Allows to export colormap data to a file.



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[y=1cm,table/col sep=comma] \pgfplotscolormaptodatafile{hot}{hot.dat} \addplot[red,mark=|] table[y index=1] {hot.dat}; \addplot[green,mark=|] table[y index=2] {hot.dat}; \addplot[blue,mark=|] table[y index=3] {hot.dat}; \end{axis} \end{tikzpicture}



copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[y=1cm,table/col sep=comma] \pgfplotscolormaptodatafile{viridis}{viridis.dat} \addplot[red] table[y index=1] {viridis.dat}; \addplot[green] table[y index=2] {viridis.dat}; \addplot[blue] table[y index=3] {viridis.dat}; \end{axis} \end{tikzpicture}

Valid options are

/pgfplots//pgfplots/colormap/output each nth=num (initially 1) ¶

Allows to downsample the color map by writing only each num’s entry.

/pgfplots//pgfplots/colormap/output format=cvs|native (initially csv) ¶

The choice csv generates a CSV file where the first column is the offset of the color map and all following are the color components.

The choice native generates a TeX file which can be \input in order to define the colormap for use in pgfplots.

/pgfplots/colormap access/extra interval width={fraction} (initially h) ¶

Attention:

this key is supposed to be a technical part of the implementation. You may want to skip its documentation as it typically works out of the box. You only need to respect sample for.

This key applies only to colormap access=piecewise constant: it ensures that each color in the colormap receives its own interval. This ensures that each color is actually visible in the output. Thus, each provided color resembles an interval. This is different from colormap access=map where each provided color resembles an interval boundary.

Normally, pgfplots, activates this feature if and only if it has automatically computed positions. Consequently, the following example implicitly uses extra interval width=h (the default):

As soon as you provide positions manually, pgfplots defaults to extra interval width=0 in order to respect the input settings:

Note that the positions make up the input nodes for the colormap with the consequence that the last color is unused; it merely serves as interval boundary. In this context, “provide manually” means positions in round braces or a non-empty value of target pos.

In order to override the input positions and get an extra interval, you have to set the option:

Note how the extra interval of colormap access=const modifies the positions of the colormap definition: they are all shifted to the left and the chosen input positions are scaled accordingly. This becomes more apparent in the following example. First, we generate a colormap with the default settings which disables the extra interval, but omits the last (brightest) color:

Next, we explicitly enable the extra interval and see that the input positions are scaled to the left. Note that they keep their relative distances, but the last color of the colormap is finally visible:

The last example relies on of colormap and a sampling procedure. In this context, pgfplots offers sample for=const which results in the expected look:

Please refer to the documentation of sample for.

The default width of this interval is the mesh width of the color map (the value h), i.e. it will always be as large as the other intervals. If the color map has nonuniform distances, the smallest encountered mesh width is used for the extra interval. This can be seen if we omit the key and add some artificial color at the right end manually – we only need to ensure that the rightmost interval has the correct length. In our example above, the smallest mesh width is \(100\), so we can generate an equivalent result by means of

As already mentioned, colormap access=const ignores the rightmost color (“red”). Note that this approach is almost the same as the internal implementation of sample for=const.

The value of colormap access/extra interval width=fraction can also be used to customize the width of the artificial interval: it is a fraction of the total width and accepted values are \(0\le \)fraction\(\le 0.9\) where \(0\) disables the extra interval and \(0.9\) corresponds to \(90\%\) of the resulting width. Any non-\(0\) value for fraction creates an extra interval and its width is fraction percent of the entire color map. The special magic value colormap access/extra interval width=h will use the colormap’s mesh width. If the colormap has nonuniform distances, it will use the smallest encountered mesh width. This is the default.

Note that fraction is a property of the colormap. The key defines it for the current color map only. Defining a new colormap uses the default width for the new colormap (but keeps the configured value for the old colormap).

Attention:

at the time of this writing, uniform color maps only support the default interval width ‘h’ and fraction\(=0\), further customization is only possible for nonuniform color maps. If you ever need to work around this limitation, you should file a feature requests and move one of your color position until the colormap becomes a nonuniform colormap.

/pgfplots/color of colormap=value

/pgfplots/color of colormap=value of colormap name

[See also Section 4.7.6.4 on page (page for section 4.7.6.4) for how to employ this within colormap definitions]

Defines the TikZ color to be the value of colormap name. If colormap name is omitted, the value of colormap name is evaluated (i.e. the current colormap is used).

The argument value is expected to be a number in the range \([0,1000]\) where \(0\) resembles the lower end of the color map and \(1000\) the upper end.

Current colormap (hot):



copy \tikz\fill [color of colormap={800}, thick, draw=.!60!black] (0,0) -- (1,1) -- (2,1) circle (10pt);

The key computes the requested color and calls color=.. Keep in mind that the magic color name ‘.’ always reflects the “current color”, i.e. the result of color=some color. Also keep in mind that color is a TikZ command which merely defines the color, you also have to provide one of ‘draw’ or ‘fill’ such that it has an effect. Since ‘.’ is a normal color, we can write draw=.!60!black to combine it with another color.

viridis:



copy \pgfplotsset{colormap name=viridis} \tikz\fill [color of colormap={1000},thick, draw=.!60!black] (0,0) -- (1,1) -- (2,1) circle (10pt);



copy \tikz\fill [color of colormap={300 of viridis},thick, draw=.!60!black] (0,0) -- (1,1) -- (2,1) circle (10pt);

The argument colormap name is either a valid argument of colormap name or a style name like colormap/cool:

cool:



copy \tikz\fill [color of colormap={300 of colormap/cool}, thick,draw=.!60!black] (0,0) -- (1,1) -- (2,1) circle (10pt);

This last syntax allows to evaluate colormaps lazily. However, if you have many references to the same colormap, it makes sense to write \pgfplotsset{colormap/cool} first followed by many references to color of colormap={... of cool} in order to avoid unnecessary lazy evaluations.

It is possible to write lots of invocations without an explicit colormap name, i.e. lots of invocations of sorts color of colormap=value. They will all use the colormap name which is active at that time.

Note that there are actually keys with two key prefixes: /pgfplots/color of colormap and an alias /tikz/color of colormap. This allows to use the keys both for plain TikZ graphics and for pgfplots.

See also colormap access=map.

/pgfplots/index of colormap=index

/pgfplots/index of colormap=index of colormap name

[See also Section 4.7.6.4 on page (page for section 4.7.6.4) for how to employ this within colormap definitions]

A variant of color of colormap which accesses the colormap name by index. Consequently, the argument index is an integer number in the range \(0,\dotsc ,N-1\) where \(N\) is the number of colors which define the colormap name. A index outside of this range is automatically clipped to the upper bound.



copy \pgfplotsset{colormap/jet} \foreach \i in { 0,...,\pgfplotscolormaplastindexof{jet} }{ \tikz\fill [ index of colormap={\i of jet}, thick, draw=.!60!black, ] (0,0) rectangle (10pt,6pt); }

\pgfplotscolormapsizeof{colormap name} ¶

Expands to the number of colors which make up colormap name.

If the argument colormap name is an unknown colormap, it expands to \(0\).

\pgfplotscolormaplastindexof{colormap name} ¶

Expands to the last index of colormap name, i.e. it is a convenience method to access \(N-1\).

If the argument colormap name is an unknown colormap, it expands to \(-1\).

See also colormap access=direct.

/pgfplots/const color of colormap=value

/pgfplots/const color of colormap=value of colormap name

[See also Section 4.7.6.4 on page (page for section 4.7.6.4) for how to employ this within colormap definitions]

Defines the TikZ color to be the value of colormap name. If colormap name is omitted, the value of colormap name is evaluated (i.e. the current colormap is used).

This key is almost the same as color of colormap, except that it uses colormap access=piecewise constant in order to determine the interpolated value.

/pgfplots/cycle list={list} ¶

/pgfplots/cycle list name={name} ¶

Allows to specify a list of plot specifications which will be used for each \addplot command without explicit plot specification. Thus, the currently active cycle list will be used if you write either \addplot+[keys] ...; or if you don’t use square brackets as in \addplot[explicit plot specification] ...;.

The list element with index \(i\) will be chosen where \(i\) is the index of the current \addplot command (see also the cycle list shift key which allows to use \(i+n\) instead). This indexing does also include plot commands which don’t use the cycle list.

There are several possibilities to change the currently active cycle list:

4.7.7.1Predefined Lists¶

Use one of the predefined lists,⁴⁰

• • color (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=color, ] \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

• • exotic (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=exotic, ] \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

• • black white (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=black white, ] \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

• • mark list (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=mark list, ] \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

  The mark list always employs the current color, but it doesn’t define one (the \addplot+ statement explicitly sets the current color to blue).

  The mark list is especially useful in conjunction with cycle multi list which allows to combine it with other lists (for example linestyles or a list of colors).

• • mark list* A list containing only markers. In contrast to mark list, all these markers are filled. They are defined as (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=mark list*, ] \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \addplot+ [blue] coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

  Similar to mark list, the mark list* always employs the current color, but it doesn’t define one (see above for the \addplot+).

• • color list (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=color list, ] \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

  The cycle list name=color choice also employs markers whereas color list uses only colors.

• • linestyles (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=linestyles, ] \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

• • linestyles* contains more dotted line styles than linestyles (from top to bottom)

  

  copy % Preamble: \pgfplotsset{width=7cm,compat=1.18} \begin{tikzpicture} \begin{axis}[ stack plots=y,stack dir=minus, cycle list name=linestyles*, ] \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \addplot coordinates {(0,1) (0.5,1) (1,1)}; \end{axis} \end{tikzpicture}

• • auto The cycle list name=auto always denotes the most recently used cycle list activated by cycle list or cycle list name.

The definitions of all predefined cycle lists follow (see the end of this paragraph for a syntax description).