I publish a lot for science journals and some of them are very strict on the number of pages your manuscript will be, as usual the final version of the articles is double column but for review and stuff like that it is always ez to compile as single column.

My question is this, Is there a way to compile just one .tex file and get a Single column pdf and a double column one?

I read that this could be achieve by use of \write18 but I don't understand at all how

I guess must be something like this:

\documentclass[layout=onecolumn]{SomeJournalStyle}
all the preamble

\write18{pdflatex "2Columnversion" Master.tex some script/modifier to change column layout at preamble }

\begin{document}
.
.
.
.
.
\end{document}

-
Why not have two 'shell' files which contains the different layouts and then \inputs the same content.tex file? Then you could also compile both with a simple loop... – jon Apr 26 '14 at 0:05
I'd like to second jon's advice - this is exactly what I ended up doing. – bbarker Apr 26 '14 at 0:25
– Peter Grill Apr 26 '14 at 1:52
@jon: I used that technique in a very similar case; creating worksheets (with blank answer spaces) and answer sheets (worksheets with the correct answers filled in). – O. R. Mapper Apr 26 '14 at 11:53

I would recommend you use a command line option to specify the column type. So the MWE below with either of these two:

pdflatex doc.tex


or

pdflatex "\def\ColumnType{onecolumn}\input{doc.tex}"


will produce a single column layout, but with

pdflatex "\def\ColumnType{twocolumn}\input{doc.tex}"


will produce a two column output.

## Code:

\ifdefined\ColumnType
\else
\def\ColumnType{onecolumn}
\fi
\documentclass[\ColumnType]{article}
\usepackage{lipsum}
\begin{document}
\lipsum[1-5]
\end{document}

-
+1 I'd use arara to help with this - then you could simply run arara on the file once, and have it to the heavy lifting :) – cmhughes Apr 26 '14 at 3:54
@cmhughes: I don't know how to do that. Perhaps you should post a separate answer. – Peter Grill Apr 26 '14 at 4:55
done! hope you don't mind that I used your code for a base :) – cmhughes Apr 26 '14 at 15:25
Thanks this works – Jared Lo Apr 29 '14 at 15:45

The arara automation tool can help to automate the idea given by Peter Grill in his answer.

You simply save the following arara rule, onetwocolumns.yaml, into a directory where arara knows to find it, and then you can run

arara myfile


on the following file, and you'll receive two pdfs:

myfileONECOLUMN.pdf
myfileTWOCOLUMN.pdf


The arara directives are detailed in the rule file but, just for clarity, you can use them in any of the following ways:

% arara: onetwocolumns
% arara: onetwocolumns: {columns: onecolumn}


The default value of columns is twocolumn but you can change that as you see fit.

# myfile.tex

% arara: onetwocolumns
% arara: onetwocolumns: {columns: onecolumn}
\ifdefined\ColumnType
\else
\def\ColumnType{onecolumn}
\fi
\documentclass[\ColumnType]{article}
\usepackage{lipsum}
\begin{document}
\lipsum[1-5]
\end{document}


# onetwocolumns.yaml

!config
# Make one and two column document
# author: Chris Hughes
# last edited by: cmh, April 26th, 2014
# http://tex.stackexchange.com/questions/173532/two-pdf-versions-from-one-single-tex-file
# requires arara 3.0+
#
# Sample usage
#
# % arara: onetwocolumns
# % arara: onetwocolumns: {columns: onecolumn}
# % arara: onetwocolumns: {columns: twocolumn}
#
identifier: onetwocolumns
name: OneTwoColumns
commands:
- <arara> pdflatex "\def\ColumnType{@{columns}}\input{@{file}}"
- <arara> @{ isWindows( "cmd /c move", "mv" ) } @{getBasename(file)}.pdf @{getBasename(file)}@{columns.toUpperCase()}.pdf
arguments:
- identifier: columns
flag: <arara> @{parameters.columns}
default: twocolumn

-

I have often used a solution like Peter Grill's for things like this, but sometimes I do it this way: a 'content' file and then two 'shell' files. I find this more convenient when there are going to be many differences in how things need to be formatted in the two files. (This can be achieved quite easily with conditionals, but sometimes I just end up going this route...)

The 'content' file is pretty simple

% content.tex
This is the file that has all the actual content


and the shell files contain all the formatting directives

% shell-1col.tex
\documentclass{article}
\usepackage{...}
% other specific stuff for the one-column layout
\begin{document}
\input{content}
\end{document}


(if the two shell files will have a lot of overlap, it is probably wise to wrap that all up in a .sty file that both include)

% shell-2col.tex
\documentclass[twocolumn]{article}
\usepackage{...}
% other specific stuff for the two-column layout
\begin{document}
\input{content}
\end{document}


And compilation is easy:

for f in shell-1col.tex shell-2col.tex ; do latex \$f ; done


As for which is better, the conditional route or the two separate files route, I suspect it depends on a number of factors. My advice is try both.

-

To achieve what has been suggested in other comments - namely, using \input to read in the meat of your document into various "shell" documents that may use different styles, I ended up doing a few additional things. I have one common file that contains helpful commands used in many of my LaTeX projects. For instance, you may want to define commands that can insert an abstract.

For conditional formatting, you may also want to use something like ifthenelse. Finally, I suggest putting the common files in a subdirectory and each "shell file" project (e.g. the file with the \documentclass) in their own directories. For more on this structure, see Path to External Files in Nested \input and How to include files from a directory specified by a concatenation of text and a macro?.

-