# How can I use alignedat in subequations?

I have a set of equations inside a subequations environment. Update: I use a macro to generate long subsubscripts in \rm.

\newcommand{\foobar}{\rm foobar}
\begin{subequations}
\begin{alignat}{5}
t_{p_{\foobar}} &=& &\big\vert \big\{e \in E^\prime \vert& &\mathcal{P}_{\foobar}(e)       &&\wedge&      &\mathcal{H}_{\foobar}(e) \big\} \big\vert\\
f_{p_{\foobar}} &=& &\big\vert \big\{e \in E^\prime \vert& &\mathcal{P}_{\foobar}(e)       &&\wedge& \lnot&\mathcal{H}_{\foobar}(e) \big\} \big\vert\\
f_{n_{\foobar}} &=& &\big\vert \big\{e \in E^\prime \vert& \lnot &\mathcal{P}_{\foobar}(e) &&\wedge&      &\mathcal{H}_{\foobar}(e) \big\} \big\vert\\
t_{n_{\foobar}} &=& &\big\vert \big\{e \in E^\prime \vert& \lnot &\mathcal{P}_{\foobar}(e) &&\wedge& \lnot&\mathcal{H}_{\foobar}(e) \big\} \big\vert
\end{alignat}
\end{subequations}


These equations have a lot of unnecessary space between columns. Using gather along with alignedat, they are aligned the way I want them - without any additional horizontal space. However, this leads to the equations being assigned only a single equation number. When I use alignat instead, latex adds additional space between columns - which I don't want. How can I achieve correct alignment and correct numbering?

-
It's not recommended to use commands such as \rm in formulas; the correct way is \newcommand{\foobar}{\mathrm{foobar}} that doesn't even require braces like in \mathcal{P}_\foobar, although braces don't harm. – egreg May 19 '11 at 16:22

You could use align and \phantoms to get the desired alignment:

\documentclass{article}
\usepackage{amsmath}

\begin{document}

\begin{subequations}
\label{eq:Classification}
\begin{align}
t_{p_{x}} &= \big\vert \big\{e \in E^\prime \mid\phantom{\lnot}\mathcal{P}_{x}(e)\wedge
\phantom{\lnot}\mathcal{H}_{x}(e) \big\} \big\vert\\
f_{p_{x}} &= \big\vert \big\{e \in E^\prime \mid\phantom{\lnot}\mathcal{P}_{x}(e)  \wedge
\lnot\mathcal{H}_{x}(e) \big\} \big\vert\\
f_{n_{x}} &= \big\vert \big\{e \in E^\prime \mid\lnot \mathcal{P}_{x}(e) \wedge
\phantom{\lnot}\mathcal{H}_{x}(e) \big\} \big\vert\\
t_{n_{x}} &=\big\vert \big\{e \in E^\prime \mid\lnot \mathcal{P}_{x}(e) \wedge
\lnot\mathcal{H}_{x}(e) \big\} \big\vert
\end{align}
\end{subequations}

\end{document}


Note that I also used \mid to get some space between the set generic elements and their defining condition.

-
Thanks, this works nicely. – Jannik Jochem May 19 '11 at 13:44

Please post minimal examples. I cannot see this in the following example

\documentclass[a4paper]{memoir}
\usepackage{amsmath}
\begin{document}
\begin{subequations}
\begin{gather}
\begin{alignedat}{5}
t_{p_{x}} &=& &\big\vert \big\{e \in E^\prime \vert&
&\mathcal{P}_{x}(e)       &&\wedge&      &\mathcal{H}_{x}(e) \big\} \big\vert\\
f_{p_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& &\mathcal{P}_{x}(e)       &&\wedge& \lnot&\mathcal{H}_{x}(e) \big\} \big\vert\\
f_{n_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& \lnot  &\mathcal{P}_{x}(e) &&\wedge&      &\mathcal{H}_{x}(e) \big\} \big\vert\\
t_{n_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& \lnot &\mathcal{P}_{x}(e) &&\wedge& \lnot&\mathcal{H}_{x}(e) \big\} \big\vert
\end{alignedat}
\end{gather}
\end{subequations}
\begin{subequations}
\begin{alignat}{5}
t_{p_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& &\mathcal{P}_{x}(e)       &&\wedge&      &\mathcal{H}_{x}(e) \big\} \big\vert\\
f_{p_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& &\mathcal{P}_{x}(e)       &&\wedge& \lnot&\mathcal{H}_{x}(e) \big\} \big\vert\\
f_{n_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& \lnot &\mathcal{P}_{x}(e) &&\wedge&      &\mathcal{H}_{x}(e) \big\} \big\vert\\
t_{n_{x}} &=& &\big\vert \big\{e \in E^\prime \vert& \lnot &\mathcal{P}_{x}(e) &&\wedge& \lnot&\mathcal{H}_{x}(e) \big\} \big\vert
\end{alignat}
\end{subequations}
\end{document}

-
You are right. To "minimize" my example, I removed the \foobar indices that seem to cause the problem in the first place. I have edited my question to show the actual problem. – Jannik Jochem May 19 '11 at 13:56
@jannik I still see no problems if pasted into my example above – daleif May 19 '11 at 14:11
Okay - then there is some weird cross-interaction of a package in my thesis. Since Gonzalo's solution fixes my problem I won't do the full bisection search to determine what exactly causes the problem. But thank you for your help. – Jannik Jochem May 19 '11 at 14:44