Oh! Sorry I think this is an unreported bug in longtable
.
Longtable inserts a strut to keep a minimal separation between the text and any \hline
or the previous row.
It does this with the following code
\def\LT@startpbox#1{%
\bgroup
\let\@footnotetext\LT@p@ftntext
\setlength\hsize{#1}%
\@arrayparboxrestore
\vrule \@height \ht\@arstrutbox \@width \z@}
If you look at the equivalent code in the array
package, you find
\def\@startpbox#1{\bgroup
% \end{macrocode}
% The argument is the width of the box. This information has to be
% assigned to =\hsize=.
% Then we assain default values to several parameters used in a
% \textsf{parbox}.
% \changes{v2.3k}{1998/05/12}{Use \cs{setlength} to set \cs{hsize},
% so that the calc package can be applied here (pr/2793)}
% \begin{macrocode}
\setlength\hsize{#1}\@arrayparboxrestore
% \end{macrocode}
% Our main problem is to obtain the same distance between succeeding
% lines of the \textsf{parbox}.
% We have to remember that the distance between two \textsf{parboxes}
% should be defined by =\@arstrut=. That means that it can be
% greater than the distance in a \textsf{parbox}.
% Therefore it is not enough to set a =\@arstrut= at the
% beginning and at the end of the \textsf{parbox}. This would
% dimension the distance
% between first and second line and the distance between the two
% last lines of the \textsf{parbox} wrongly.
% To prevent this we set an invisible rule of height
% =\@arstrutbox=
% at the beginning of the \textsf{parbox}. This has no effect on the
% depth of the first line. At the end of the \textsf{parbox} we set
% analogously another invisible rule which only affects the depth
% of the last line. It is necessary to wait inserting this strut
% until the paragraph actually starts to allow for things like
% =\parindent= changes via =>{...}=.
% \changes{v2.1c}{1992/12/14}{Use `everypar to insert strut}
% \begin{macrocode}
\everypar{%
\vrule \@height \ht\@arstrutbox \@width \z@
\everypar{}}%
}
% \end{macrocode}
% \end{macro}
The important bit being the last few lines which delay adding the strut by using \everypar
.
It seems to me the same fix should have been applied to longtable
so only 20 years too late I suggest
\def\LT@startpbox#1{%
\bgroup
\color@begingroup% Omit line if package date older than 2014/10/28
\let\@footnotetext\LT@p@ftntext
\setlength\hsize{#1}%
\@arrayparboxrestore
\everypar{%
\vrule \@height \ht\@arstrutbox \@width \z@
\everypar{}}%
% \vrule \@height \ht\@arstrutbox \@width \z@
}
which then makes both your tables have the same vertical spacing.
If you do this could you change
\ProvidesPackage{longtable}
[2004/02/01 v4.11 Multi-page Table package (DPC)]
to
\ProvidesPackage{longtable}
[2004/02/01 v4.11 Multi-page Table package (DPC) + strut fix]
I suspect after all this time I can't change longtable
s default spacing, however I should add an option to fix it.
You should use split
inside equation
for this:
\documentclass{article}
\usepackage{amsmath}
\begin{document}
\begin{equation}
\begin{split}
\pi^{(1)}_{i j}
={}& -\frac{1}{4}\eta^{\lambda \sigma} ( \eta^{\mu \alpha}\eta^{\nu \beta} + \eta^{\mu \beta}\eta^{\nu \alpha} ) \biggl\{ \frac{1}{2}\delta_{\lambda 0}( \delta_{\mu i}\delta_{\nu j} + \delta_{\mu j}\delta_{\nu i} ) \biggr\} \partial_{\sigma} h_{\alpha \beta}
\\
& -\frac{1}{4}\eta^{\lambda \sigma} ( \eta^{\mu \alpha}\eta^{\nu \beta} + \eta^{\mu \beta}\eta^{\nu \alpha} ) \biggl\{ \frac{1}{2}\delta_{\sigma 0}( \delta_{\alpha i}\delta_{\beta j} + \delta_{\alpha j}\delta_{\beta i} ) \biggr\} \partial_{\lambda} h_{\mu \nu}
\\
={}& -\frac{1}{4}\partial^{\lambda} h^{\mu \nu}\delta_{\lambda 0}( \delta_{\mu i}\delta_{\nu j} + \delta_{\mu j}\delta_{\nu i} ) -\frac{1}{4}\partial^{\sigma} h^{\alpha \beta}\delta_{\sigma 0}( \delta_{\alpha i}\delta_{\beta j} + \delta_{\alpha j}\delta_{\beta i} )
\\
={}& -\frac{1}{2}\partial^0 h^{i j} - \frac{1}{2}\partial^0 h^{i j}
\\
={}& \dot{h}_{i j}
\end{split}
\end{equation}
\end{document}
The trick is to put =
on the left side, following it with {}
for ensuring correct spacing.
Avoid useless \left
and \right
. Also, no \enspace
should be used around =
.
![enter image description here](https://i.stack.imgur.com/Hgjfy.png)
The number should go in the middle, in my opinion. However, the tbtags
option to amsmath
will put it at the bottom of a split
.
Best Answer
For one-off spacing changes,
\smallskip
,\medskip
, and\bigskip
are defined for this task.After rereading the question, I'll note that these aren't 'fixed' spaces; they're rubber lengths (defined according to
\smallskipamount
, etc.) and will shrink/expand according to the amount of available vertical space when the page is being built.