Comprehensive list of tools that simplify the generation of LaTeX tables

Question

Manually hacking tables in LaTeX is one of the less funny things when preparing a document. Naturally, there are quite a few tools that promise to simplify this. In fact, I have somewhat lost track of all the tools available, and I'd like to gather your valuable experience into one big post. Some of the tools are mentioned in Tool for manipulating LaTeX tables (HTML to LaTeX), but I feel there's much more around.

I'd suggest that somebody starts an overview with an alphabetic list (so that I can accept this answer and it stays on top), and that everyone replies with a "community wiki" entry, one reply per tool, and inserts the tool into the list.

Personally, I use Excel2LaTeX, but I'm unhappy with having to fire up Windows in order to generate the LaTeX code. I'm thinking of a rewrite as a standalone Python tool using the xlrd library, but I'm not too inclined to reinvent the wheel.

Answer 1

Often, I have the use-case that I want to convert a given data table into a "suitable" LaTeX table.

Typically, my data is of numeric type and requires number formatting, perhaps alignment at a decimal point, and in most cases, it requires elementary post-processing (like quotients, differences, gradients).

Since I needed such stuff very often, I wrote some C++ scripts which generated .tex files. It was useful - but I realized very early that it is insufficient; it is just not flexible enough and -as any external tool- produces unwanted complications due to the many tools involved.

My solution is the LaTeX package pgfplotstable. It is a LaTeX table generator, i.e. it converts input data explicitly by a set of configurable rules into something like \begin{tabular}....\end{tabular}.

DISCLAIMER NOTE: I wrote the package.

Among its features are

• separation of data + format
• data in form of external data files (CSV with customizable separators) or inline tables (inside of the .tex file)
• central format definition (for example in the preamble or in form of styles)
• format numbers with the full power of LaTeX
• supports simple text columns as well
• simple support for alternating row colors (colortbl)
• simple support for standart LaTeX table packages (booktabs, longtable, colortbl, multirow, \multicolumn,...)
• can produce completely new columns containing postprocessed data, with the powerful pgf math engine
• can convert single-column output to two-column output
• is written completely in TeX (no external tools required)
• highly customizable
• has a manual with lots of examples.

You may want to inspect the link mentioned above and its examples to see if it fits your needs.

Answer 2

^{I was reluctant to publish this humble tool, but here it goes.}

excel2datatool

excel2datatool is a Java application I wrote to help me with a personal project. Since it was useful for me and I do believe it's generic enough for other purposes, I decided to make it publicly available. The repository is hosted on GitHub.

The whole idea is to read Microsoft Excel formats and generate .csv + .tex + datatool. Some might argue that this app is basically a CSV converter. Well, it is, but since I had to write a complementary .tex file with \DTLloaddb, I decided to automate steps.

The app does not rely on Microsoft Excel. It supports both xls and xlsx out of the box. The usage is quite easy, double-click excel2datatool.jar in Windows and Mac, or run java -jar excel2datatool in Linux. The code is Java 5, 6, 7 and OpenJDK 6 compliant.

Pros

• There's no need for Microsoft Excel or other office tool.
• It supports both xls and xlsx formats out of the box.
• Runs on every major operating system.
• The LaTeX file generation is optional, so you can use it as a normal CSV converter.

Cons

• I wrote the app.
• Since it doesn't rely on external applications, the payload is quite big (9.5Mb).
• The LaTeX file will contain the most basic datatool usage, with \DTLloaddb. There are no advanced commands.
• No command line interface.
• Some people might not like the drag-and-drop style.
• The application tries its best to parse the Excel file, but there are probably some more complex files which won't be correctly converted.

Answer 3

Orgmode

Orgmode is a notes/planning mode for Emacs that has a nice table feature.

http://orgmode.org/

Pros

• Column widths auto-adjust based on contents
• Behaves as a spreadsheet, including calculations, row/column insert and delete, row/column moves, rectangular selection
• Import and export of TAB or whitespace separated data
• Lots of export options (LaTeX, HTML, PDF)

Cons

• Requires Emacs which has a significant learning curve if you are not already using it
• Some advanced LaTeX table formatting is still done manually after export

Answer 4

R + xtable package

R has the package xtable which generates LaTeX output - useful if you want R results in tabular form. This is an easy way if you're working in the R environment anyway.

Free.

Pros

• At least the following table types can be converted to a nice table out of the box:

  • One-way and two-way cross tabulations
  • Data frames
  • Model estimation results

  (Anyone more familiar with xtable please append to this list.)

• Support for booktabs and longtable

• Optional enclosing in table environment with caption
• Formatting of numbers and cell contents can be specified
• Optional filtering ("sanitizing") of TeX special characters
• More flexibility in conjunction with knitr/Sweave: The definition of the table can be "plain" LaTeX code, and the contents can be generated by an R chunk

Cons

• Requires R, steep learning curve if you are not familiar with R

• No support for multicolumn and multirow cells

• Any special effects (such as row and column headers for a contingency table require programming)

See also the following question for details on using xtable:

How can I use a table generated by R in LaTeX?

The Hmisc package can also generate tabular output from R.

Answer 5

Excel2LaTeX

An Excel add-in that converts parts of a spreadsheet into equivalent LaTeX tables.

Free.

http://www.ctan.org/tex-archive/support/excel2latex/

Pros

• Supports and converts various formatting options:

  • bold/italics
  • alignment
  • cell spanning (horizontal/vertical)
  • struts
  • rotation of cell contents
  • cell border lines
  • booktabs option

• Configurable: Repeatedly convert the same ranges with one click

Cons

• Requires Windows or MacOS, and Excel to do the conversion

Disclaimer: I am the current maintainer of this add-in.

Answer 6

LaTable

A visual editor for LaTeX tables for Windows and Macintosh. It can be downloaded from CTAN.

The application provides a spreadsheet-like interface with the ability to insert and delete rows and columns, to edit the cells and to set the borders. Tables can be im- and exported as plain text or CSV files or saved in LaTable's own file format .latbl.

A (rather horribly-looking) example:

The code generated for this example table:

\begin{tabular}{l|ll|l}
\cline{3-3}
first column & \multicolumn{1}{l|}{second column} & third column & fourth column \\ 
\hline
\multicolumn{1}{c|}{1} & \multicolumn{1}{l|}{2} & 3 & 4 \\ 
\cline{3-3}
5 & 6 & \multicolumn{1}{r}{7} & 8 \\ 
\cline{1-2}\cline{4-4}
9 & \multicolumn{1}{c}{10} & 11 & \multicolumn{1}{l|}{12} \\ 
\cline{4-4}
\end{tabular}

And this is the compiled LaTeX result for comparison:

Answer 7

Gnumeric

My prefered solution.

EDIT:

• You can save the spreadsheet as a full *.tex document (or one which can be "inserted" into an existing one, instructions for this are given as % comments %).
• Alternatively, (my personal preference) gnumeric allows you to save the spreadsheet as a "table fragment" which will simply give you the plain-vanilla TeX-table :)

See http://projects.gnome.org/gnumeric/doc/file-format-latex.shtml for some further details.

Answer 8

LyX

LyX has a relatively decent table editor. By selecting View->View Source, one can directly see the resulting LaTeX code and easily copy&paste (and also learn) it from there. Furthermore, it is easy to typeset the result via LaTeX to get an impression on the actual look of the table. Other benefits are built-in support for booktabs and longtable.

The immediate representation of the resulting LaTeX source helped me a lot to actually learn the whistles and bells about LaTeX tabulars. Today I am fluent enough to use LyX only very seldom for this purpose.

Answer 9

Calc2LaTeX

For the sake of completeness: You know, there is the spreadsheet programm of the OpenOffice / LibreOffice suite, simply called »calc«. It runs on Linux, Windows and probably on Mac, and it imports Excel spreadsheets quite nicely, if you haven't done more than using tables.

From calc to LaTeX there is an add to calc: calc2Latex

Works nicely, I've been using it for years.

Answer 10

csv2latex

For Mac users there is the script csv2latex which allows you to cut and paste spreadsheet cells from Excel, Number or Calc and paste them into your LaTeX source in a variety of popular table formats.

If you use TeXShop as your editor, this functionality is built into the program: there is a Paste Spreadsheet Cells macro in the Macros menu. For other Applescript-aware Mac editors, you can install the script manually.

Disclaimer: I am the author.

Answer 11

Emacs align-current and rectangles

OK, this isn't really much of an answer, but if you already have a table in LaTeX and you just want to edit it a bit, it can get a bit messy. If you use emacs then the following tools are super helpful.

First, align-current tries to line up your &s so that your data is more or less in columns. You can then use emacs' C-x r k to cut rectangles out, and C-x r y to paste rectangles back into your table.

Here are some links:

• Blog post about the above trick
• Another post about rectangles
• Emacs manual on rectangles

Answer 12

Numbers2LaTeX

A quick hack from long time ago that converts a table in numbers (for Mac OS X) to a LaTeX table: https://gist.github.com/386384

Answer 13

Warning: what's below is more of a tongue-in-cheek answer, but not without a shred of truth in it!

\halign & \valign

Pro

• It's what you'd eventually get with the other packages (most likely) anyways
• format agnostic – it's a TeX primitive
• Once you grok it, you'll be able to produce typographically beautiful tables more easily, and, on the other hand,
• creating typographically ugly tables is difficult

Contra

• there's a learning curve
• creating typographically ugly tables is difficult ;-)

---

To expand a little, to think about what one is doing when inputting tabular data is that, if, for example, one is working on a spreadsheet program, what you are mentally doing is that every time you want to express that you want to move on to the next horizontal cell you press tab, and every time you want to move to the next row you press enter.

Now, if you just input & instead of tab, and \cr instead of enter, you can skip all the intermediary steps. That is, converting the spreadsheet program generated chunk into tabular (or whatever), and then expanding that out to a \halign.

Similarly, by the time that you've navigated with your mouse through the various menus in the spreadsheet program to express that you want a particular cell to span, say, three columns (this would almost always mean that you are working on a “header” cell), you would have already written \multispan3.

I can already see that I'm writing myself into a corner here, what with the various arguments for cell background colors, cells spanning multiple rows, cells containing vertical mode material, and what have you. So I'm just going to say a sweeping statement about all those things being “bad typography”. ;-).

Hold on, there's a lynch mob knocking on my door…

What I mean is that if you find yourself doing any of the above, more often than not, chances are that you are complicating the information you want to communicate with your table. And that's basically what the word “bad typography” means.

So one strong argument still remains: to visually see the structure of the table, be it in raw text (like in the orgmode answer, vims equivalent, or in the various Markdown extensions), or in graphical programs (like in spreadsheet programs).

This could again be just a s/<colsepchar>/\&/g and s/<rowsepchar>/\\cr/, in whatever chunk you get out from the program of your choice. Like, not a big deal at all. Or alternatively on the TeX side of things, some \catcode madness could take place. Given that the table is simple enough, i.e., not complicating the information you want to communicate with it.

So, given this context and background, I hope my initial joke's not lost on anyone and this answer could happily live in this question! ;-)

Answer 14

Kile

http://docs.kde.org/stable/en/extragear-office/kile/wizard_array.html

http://docs.kde.org/stable/en/extragear-office/kile/wizard_array.html

Arrays and tabulars

One of the most boring jobs one can do in LATEX is to write a matrix or a tabular environment. One has to keep track of all the elements, ensure that the environment is well formed, and that all things are where they are supposed to be. Good indentation helps, but there is a simpler way: using Kile's Wizard → Array or Wizard → Tabular menu entries. You will then have a matrix-formed input form that you can easily fill in with your entries. This dialog also offers some options to typeset the tabular material.

Answer 15

Direct spreadsheet editing

I've been handling data tables from Excel into LaTeX for lab reports for some while and I have tried several tools. I would like to make the case that unless you consistently require formatting of the same style, then quite often the simplest answer is to direct copy-paste from your spreadsheet into LaTeX with inserted columns including whatever separators you want.

This is particularly useful if you need extra mark-up between columns, such as $ ... $ signs or extra mathematical stuff such as units and the like.

This is arguably best for producing tableaux rather than formal scientific tables, and it does have the disadvantage that it inextricably links up data and formatting. On the other hand, that gives you a lot of freedom for data-dependent formatting. My most recent use of this was the following table, for my dissertation:

I can really not think of other ways of doing this - and particularly of ways that would later on enable me to easily change one of the functions displayed. This is of course not a scientific table, but this tableau is I think the clearest such presentation of these functions. (Spreadsheet and source code available on request.)

Answer 16

Did someone miss:

tablas

Some dummy text to fill...

Answer 17

As you know you can also get a snapshot of your table and add the image file to your file. OK, OK! you are right this is not a good solution for all cases but use it as last resort. In such situation http://getgreenshot.org/ is your assistant.