pgfplots使用手册：创建2D/3D图表

需积分: 32 56 浏览量更新于2024-07-15 收藏 12.34MB PDF 举报

"LaTeX pgfplots使用指南" 在LaTeX中创建高质量的2D和3D图表时，pgfplots是一个强大的工具。该包由Christian Feuersänger开发，旨在提供一个用户友好的接口，用于在LaTeX环境中直接绘制各种类型的图形。通过pgfplots，用户只需提供坐标轴标签、图例内容以及一个或多个图形的坐标数据，包就会自动处理轴的缩放、计算对数、设置轴刻度，并绘制出图形。它支持线性图、散点图、阶梯图、条形图、面积图、网状图和曲面图、等高线图、向量场图、直方图、极坐标图、三元图、史密斯图等多种图形类型。文档内容包括以下几个部分： 1. **引言**：这部分通常会介绍pgfplots包的基本概念，解释其设计目标和用途，以及为何在LaTeX中使用它来创建图形是理想的选择。 2. **关于pgfplots：预备知识**：这一章节详细介绍了pgfplots的组成部分，如命令、环境和选项。同时，还会包含升级注意事项，比如新版本中增加的可选功能和可能需要关注的旧功能变化。 - **新可选功能**：这部分会列出新版本中引入的特性，帮助用户了解如何利用它们来增强图形的表达力。 - **可能需要注意的旧功能**：可能会提到某些旧功能在新版本中的变化，以便用户在更新时避免出现不兼容问题。 3. **团队**：简单介绍pgfplots的开发团队成员，以及他们的贡献。 4. **致谢**：感谢那些对pgfplots发展做出贡献的人。 5. **安装与先决条件**：指导用户如何安装pgfplots包，以及需要哪些前提条件。 - **许可**：说明pgfplots的许可协议，确保用户了解使用该包的法律条款。 - **前提条件**：列出使用pgfplots前需要安装的其他LaTeX包，如pgf/TikZ，这是pgfplots的基础。 pgfplots基于Till Tantau的pgf/TikZ包，这意味着它能充分利用TikZ的强大绘图能力。用户可以通过LaTeX语法直接控制图形的每一个细节，实现高度定制化。通过学习和掌握pgfplots，用户可以在科研报告、学术论文或者任何需要精美图表的文档中，轻松地生成专业级别的图形。

Chapter 3

Step-by-Step Tutorials

3.1 Introduction

Visualization of data is often necessary and convenient in order to analyze and communicate results of

research, theses, or perhaps just results.

pgfplots is a visualization tool. The motivation for pgfplots is that you as end-user provide the data

and the descriptions as input, and pgfplots takes care of rest such as choosing suitable scaling factors,

scaling to a prescribed target dimension, choosing a good displayed range, assigning tick positions, drawing

an axis with descriptions placed at appropriate places.

pgfplots is a solution for an old problem of visualization in L

X: its descriptions use the same fonts as

the embedding text, with exactly the same font sizes. Its direct embedding in L

X makes the use of L

X’s

powerful math mode as easy as possible: for any kind of axis descriptions up to user-deﬁned annotations. It

features do cument-wide line-styles, color schemes, markers. . . all that makes up consistency.

pgfplots oﬀers high-quality. At the same time, it is an embedded solution: it is largely independent of

3rd party tools, although it features import functions to beneﬁt from available tools.

Its main goal is: you provide your data and your descriptions – and pgfplots runs without more input.

If you want, you can customize what you want.

3.2 Solving a Real Use Case: Function Visualization

In this section, we assume that you want to visualize two functions. The ﬁrst function is given by means of

a data table. The second function is given by means of a math expression. We would like to place the two

results side by side, and we would like to have “proper” alignment (whatever that means).

As motivated, we have one data table. Let us assume that it is as shown below.

x_0 f(x)

# some comment line

3.16693000e-05 -4.00001451e+00

1.00816962e-03 -3.08781504e+00

1.98466995e-03 -2.88058811e+00

2.96117027e-03 -2.75205040e+00

3.93767059e-03 -2.65736805e+00

4.91417091e-03 -2.58181091e+00

5.89067124e-03 -2.51862689e+00

...

9.89226496e-01 2.29825980e+00

9.90202997e-01 2.33403276e+00

9.91179497e-01 2.37306821e+00

9.92155997e-01 2.41609413e+00

9.93132498e-01 2.46412019e+00

9.94108998e-01 2.51860712e+00

9.95085498e-01 2.58178769e+00

9.96061999e-01 2.65733975e+00

9.97038499e-01 2.75201383e+00

9.98014999e-01 2.88053559e+00

9.98991500e-01 3.08771757e+00

9.99968000e-01 3.99755546e+00

Note that parts of the data ﬁle have been omitted here because it is a bit lengthy. The data ﬁle

3.2. SOLVING A REAL USE CASE: FUNCTION VISUALIZATION 17

(and all others referenced in this manual) are shipped with pgfplots; you can ﬁnd them in the subfolder

doc/latex/pgfplots/plotdata.

3.2.1 Getting the Data Into T

Our ﬁrst step is to get the data table into pgfplots. In addition, we want axis descriptions for the x and

y axes and a title on top of the plot.

Our ﬁrst version looks like

0 0.2 0.4 0.6 0.8 1

−4

−2

Inv. cum. normal

% Preamble: \pgfplotsset{width=7cm,compat=1.15}

%\documentclass{article}

%\usepackage{pgfplots}

%\pgfplotsset{compat=1.5}

%\begin{document}

\begin{tikzpicture}

\begin{axis}[

title=Inv. cum. normal,

xlabel={$x$},

ylabel={$y$},

]

\addplot [blue] table {invcum.dat};

\end{axis}

\end{tikzpicture}

%\end{document}

The co de listing already shows a couple of important aspects:

1. As usual in L

X, you include the package using \usepackage{pgfplots}.

2. Not so common is \pgfplotsset{compat=1.5} .

A statement like this should always be used in order to (a) beneﬁt from a more or less recent feature

set and (b) avoid changes to your picture if you recompile it with a later version of pgfplots.

Note that pgfplots will generate some suggested value into your logﬁle (since 1.6.1). The minimum

suggested version is \pgfplotsset{compat=1.3} as this has great eﬀect on the positioning of axis

labels.

3. pgfplots relies on TikZ and pgf. You can say it is a “third party package” on top of TikZ/pgf.

Consequently, we have to write each pgfplots graph into a Tik Z picture, hence the picture environ-

ment given by \begin{tikzpicture} ... \end{tikzpicture}.

4. Each axis in pgfplots is written into a separate environment. In our case, we chose \begin{axis}

... \end{axis} as this is the environment for a normal axis.

There are more axis environments (like the \begin{loglogaxis} ... \end{loglogaxis} environ-

ment for logarithmic axes).

Although pgfplots runs with default options, it accepts keys. Lots of keys. Typically, you provide

all keys which you “want to have” in square brackets “somewhere” and ignore all other keys.

Of course, the main diﬃculty is to get an overview over the available keys and to ﬁnd out how to use

them. This reference manual and esp ecially its Section 4 has been designed for online browsing: it

contains hundreds of cross-referenced examples. Opening the manual in a PDF viewer and searching

it for keywords will hopefully jump to a good match from which you can jump to the reference section

(for example about tick labels , tick positions, plot handlers etc.). It is (and will always be) the most

reliable source of detail information about all keys.

Speaking about the reference manual: note that most PDF viewers also have a function to “jump back

to the page before you clicked on a hyperlink” (for Acrobat Reader, open the menu View / Toolbars

/ More Tools and activate the “Previous View” and “Next View” buttons which are under “Page

Navigation Toolbar”).

Note that the code listing contains two sets of keys: the ﬁrst is after \begin{axis}[ ... ] and

the second right after \addplot[...]. Note furthermore that the option list after the axis has been

18 CHAPTER 3. STEP-BY-STEP TUTORIALS

indented: each option is on a separate line, and each line has a tab stop as ﬁrst character. This is a

good practice. Another good practice is to place a comma after the last option (in our case, after the

value for ylabel). This allows to add more keys easily – and you won’t forget the comma. It does

not hurt at all. The second “set” of keys after \addplot shows that indentation and trailing comma

a really just a best practice: we simply said \addplot[blue], meaning that the plot will be placed in

blue color, without any plot mark. Of course, once another option would be added here, it would be

best to switch to indentation and trailing comma:

\addplot[

blue,

mark=*,

]

table {invcum.dat};

5. Inside of an axis, pgfplots accepts an \addplot ... ; statement (note the ﬁnal semicolon).

In our case, we use \addplot table: it loads a table from a ﬁle and plots the ﬁrst two columns.

There are, however, more input methods. The most important available inputs methods are \addplot

expression (which samples from some mathematical expression) and \addplot table (loads data

from tables), and a combination of both which is also supported by \addplot table (loads data from

tables and applies mathematical expressions). Besides those tools which rely only on built-in methods,

there is also an option to calculated data using external to ols: \addplot gnuplot which uses gnuplot

as “desktop calculator” and imports numerical data, \addplot shell (which can load table data from

any system call), and the special \addplot graphics tool which loads an image together with meta

data and draws only the associated axis.

In our axis, we ﬁnd a couple of tokens: the ﬁrst is the m andatory \addplot token. It “starts” a further

plot. The second is the option list for that plot, which is delimited by square brackets (see also the

notes about best practices above). The name “option list” indicates that this list can be empty. It

can also be omitted completely in which case pgfplots will choose an option list from its current

cycle list (more about that in a diﬀerent lecture). The next token is the keyword “table”. It tells

pgfplots that table data follows. The keyword “table” also accepts an option list (for example, to

choose columns, to deﬁne a diﬀerent col sep or row sep or to provide some math expression which

is applied to each row). More on that in a diﬀerent lecture. The next token is {invcum.dat}: an

argument in curly braces which provides the table data. This argument is interpreted by “plot table”.

Other input types would expect diﬀerent types of arguments. In our case, the curly braces contain a

ﬁle name. Plot table expects either a ﬁle name as in our case or a so-called “inline table”. An inline

table means that you would simply insert the contents of your ﬁle inside of the curly braces. In our

case, the table is too long to be inserted into the argument, so we place it into a separate ﬁle. Finally,

the last (mandatory!) token is a semicolon. It terminates the \addplot statement.

6. Axis descriptions can be added using the keys title, xlabel, ylabel as we have in our example

listing.

pgfplots accepts lots of keys – and sometimes it is the art of ﬁnding just the one that you were looking

for. Hop efully, a search through the table of contents of the reference manual and/or a keyword search

through the entire reference manual will show a hit.

3.2.2 Fine-Tuning of the First Picture

While looking at the result of Section 3.2.1, we decide that we want to change something. First, we decide

that the open ends on the left and on the right are disturbing (perhaps we have a strange taste – or perhaps

we know in advance that the underlying function is not limited to any interval). Anyway, we would like to

show it only in the y interval from −3 to +3.

We can do so as follows:

3.2. SOLVING A REAL USE CASE: FUNCTION VISUALIZATION 19

0 0.2 0.4 0.6 0.8 1

−2

Inv. cum. normal

% Preamble: \pgfplotsset{width=7cm,compat=1.15}

\begin{tikzpicture}

\begin{axis}[

title=Inv. cum. normal,

xlabel={$x$},

ylabel={$y$},

ymin=-3, ymax=3,

minor y tick num=1,

]

\addplot [blue] table {invcum.dat};

\end{axis}

\end{tikzpicture}

We added three more options to the option list of the axis. The ﬁrst pair is ymin=-3 and ymax=3. Note

that we have placed them on the same line although we said the each should be on a separate line. Line

breaks are really optional; and in this case, the two options appear to belong together. They deﬁne the

display limits. Display limits deﬁne the “window” of the axis. Note that any \addplot statements might

have more data (as in our case). They would still generate graphics for their complete set of data points!

The keys ymin,ymax,xmin,xmax control only the visible part, i.e. the axis range. Everything else is clipped

away (by default). The third new option is minor y tick num=1 which allows to customize minor ticks.

Note that minor ticks are only displayed if the major ticks have the same distance as in our example.

Note that we could also have modiﬁed the width and/or height of the ﬁgure (the keys have these

names). We could also have used one of the predeﬁned styles like tiny or small in order to modify not

just the graphics, but also use diﬀerent fonts for the descriptions. We could also have chosen to adjust the

unspeciﬁed limits: either by ﬁxing them explicitly (as we did for y above) or by modifying the enlargelimits

key (for example using enlargelimits=false).

We are now satisﬁed with the ﬁrst picture and we would like to add the second one.

3.2.3 Adding the Second Picture with a Diﬀerent Plot

As motivated, our goal is to have two separate axes placed side by side. The second axis should show

a function given as math expression. More precisely, we want to show the density function of a normal

distribution here (which is just a special math expression).

We simply start a new tikzpicture and insert a new axis environment (perhaps by copy–pasting our

existing one). The \addplot command is diﬀerent, though:

−2 0 2

·10

−3

100

200

300

400

% Preamble: \pgfplotsset{width=7cm,compat=1.15}

\begin{tikzpicture}

\begin{axis}[

]

% density of Normal distribution:

\addplot [

red,

domain=-3e-3:3e-3,

samples=201,

]

{exp(-x^2 / (2e-3^2)) / (1e-3 * sqrt(2*pi))};

\end{axis}

\end{tikzpicture}

We see that it has an axis environment with an empty option list. This is quite acceptable: after all, it

is to be expected that we will add options eventually. Even if we don’t: it does not hurt. Then, we ﬁnd the

expected \addplot statement. As already explained, \addplot statements initiate a new plot. It is followed

by an (optional) option list, then by some keyword which identiﬁes the way input coordinates are provided,

then arguments, and ﬁnally a semicolon. In our case, we ﬁnd an option list which results in a red plot.

The two keys domain and samples control how our math expression is to be evaluated: domain deﬁnes the

sampling interval in the form a:b and samples=N expects the number of samples inserted into the sampling

20 CHAPTER 3. STEP-BY-STEP TUTORIALS

interval. Note that domain merely controls which samples are taken; it is independent of the displayed axis

range (and both can diﬀer signiﬁcantly). If the keyword deﬁning how coordinates are provided is missing,

pgfplots assumes that the next argument is a math expression. Consequently, the ﬁrst token after the

option list is a math expression in curly braces. We entered the density function of a normal distribution

here (compare Wikipedia).

Note that the axis has an axis multiplier: the x tick labels have been chosen to be −2, 0, and 2 and an

extra x tick scale label of the form ‘·10

−3

’. These tick scale labels are quite convenient are are automatically

deduced from the input data. We will see an example with the eﬀects of scaled x ticks=false at the end

of this tutorial.

Inside of the math expression, you can use a lot of math functions like exp, sin, cos, sqrt, you

can use exponents using the a^b syntax, and the sampling variable is x by default. Note, however,

that trigonometric functions operate on degrees by default! If you need to sample the sinus func-

tion, you can use \addplot[domain=0:360] {sin(x)};. This is quite uncommon. You can also use

\addplot[domain=0:2*pi] {sin(deg(x)};. This samples radians (which is more common). But since

the math parser expects degrees, we have to convert x to degrees ﬁrst using the deg() function. See also

trig format plots=rad. The math parser is written in T

X (it does not need any third-party tool). It

supports the full range of a double precision number, even though the accuracy is about that of a single

precision number. This is typically more than suﬃcient to sample any function accurately. If you ever en-

counter diﬃculties with precision, you can still resort to \addplot gnuplot in order to invoke the external

tool gnuplot as “coordinate calculator”.

The experienced reader might wonder about constant math expressions domain=-3e-3:3e-3, 2e-3^2,

and 1e-3 rather than some variable name like ‘mu’ or ‘sigma’. This is actually a matter of taste: both is

supported and we will switch to variable names in the next listing.

The main part of our step here is still to be done: we wanted to place two ﬁgures side by side. This can

be done as follows:

0 0.2 0.4 0.6 0.8 1

−2

Inv. cum. normal

−2 0 2

·10

−3

200

400

剩余565页未读，继续阅读

qq_36682458

粉丝: 0
资源: 2

pgfplots使用手册：创建2D/3D图表

matlab2dplot2datafi​le:使用 LaTeX 中的 TikZ/pgfplots 将大型绘图数据集从 matlab 绘图压缩到数据文件 (txt/csv/dat) 以进行数据可视化-matlab开发

数学论文书籍排版Latex使用指南

PGFPlots绘图指南： LaTeX论文作图神器

latex 插图指南

LATEX中文使用手册

latex中文使用手册

LaTeX插图指南（tex文件）

LaTeX2e插图指南.pdf

pgfplots1.11：LaTeX 2D/3D绘图指南

掌握LaTeX绘图：TikZ与pgfplots包使用手册

最新资源

matlab2dplot2datafile:使用 LaTeX 中的 TikZ/pgfplots 将大型绘图数据集从 matlab 绘图压缩到数据文件 (txt/csv/dat) 以进行数据可视化-matlab开发