In the MWE example below I have the values specified twice, but would prefer not to have to do that. Furthermore, it is possible to allow PGFplots to automatically select these end points so I would have to iterate and modify the settings manually after seeing what PGFplots had selected. There must be some way to access these variables form within the axis
environment:
\documentclass{standalone}
\usepackage{pgfplots}
\begin{document}
\begin{tikzpicture}
\begin{axis}[xmin=-1.5, xmax=1.5, ymin=-1, ymax=3]
\addplot[mark=none, domain=-2.5:2.5, thick, red] ({x},{x*x-0.5});%
% Would prefer to be able to specify something like this:
%\addplot [mark=none, blue, ultra thick] coordinates {(\xmin,\ymin) (\xmin,\ymax)};
% But until I can access those values, do it manually:
\addplot [mark=none, blue, ultra thick] coordinates {(-1.5,-1) (-1.5,3)};
\end{axis}
\end{tikzpicture}
\end{document}
In case there is an interest in why I want to do this: I am trying to automate labelling intersections with the lines x=0 and y=0 within PGFplots as discussed in How to specify a name path for the axis in PGFplots. However, there are issues with that approach. One solution (albeit it is a bit of a hack) is to draw the lines x=0, and y=0 with draw=none
option and then find the intersection with that. If I can access the min/max values of the x and y axis then I can label a phantom line and automate the location of the intersections via Intersections in PGFplots