pgfplots: Punkte eines plots benennen und in einem anderen plot verwenden

\documentclass[margin=5mm, tikz]{standalone}
\usetikzlibrary{angles} 
\usepackage{pgfplots}
\usepgfplotslibrary{patchplots}
\pgfplotsset{compat=1.13}

\begin{document}

\begin{tikzpicture}
\begin{axis}[hide axis, 
colormap={Kantenfarbe}{color=(gray) color=(gray)},
]

% Zeichnung der Punkte =====================
\addplot+[nodes near coords=\coordindex, 
nodes near coords style={%%
name=p\coordindex,
path picture={%
\coordinate[] (P\coordindex) at (p\coordindex.south);
},%
},%%
]
table[header=true, x index=1, y index=2, row sep=\\] {
Nr x y  \\
1 1 3             \\%0
2 0 2              \\%1
3 2 2              \\%2
4 2 1              \\%3
};

% Tabellarische Zeichnung von Annotationen =================

\addplot[no marks,
nodes near coords={},
%
% Versuch ......
visualization depends on={value \thisrowno{0} \as \punktnummer},
nodes near coords style={minimum size=9cm,
path picture={
\draw[red, thick] (P\punktnummer) circle[radius=7mm] node[left]{geht nicht};
}},
]
table[header=true, x expr =0, y expr=0]{
Nr.
3
2
};

\end{axis}

% Test 
\draw[brown] (P0) circle[radius=5mm] node[right]{geht};
\draw[brown] (P1) circle[radius=5mm] node[right]{geht};
\draw[brown] (P1) -- (P3) node[midway, right]{geht auch};

\end{tikzpicture}
\end{document}

\documentclass[margin=5mm, tikz]{standalone}
\usetikzlibrary{angles,calc} 
\usepackage{pgfplots}
\usepgfplotslibrary{patchplots}
\pgfplotsset{compat=1.13}

\begin{document}

\begin{tikzpicture}[get angle
path/.code={\pgfmathtruncatemacro{\ang}{abs(\n2-\n1)}
\ifnum\ang>180
\xdef#1{1}
\else
\xdef#1{0}
\fi
\typeout{\n1,\n2,\ang}}]
\begin{axis}[hide axis, 
colormap={Kantenfarbe}{color=(gray) color=(gray)},
]

% Zeichnung der Punkte =====================
\addplot+[nodes near coords=\coordindex, 
nodes near coords style={%%
name=p\coordindex,
path picture={%
\coordinate[] (P\coordindex) at (p\coordindex.south);
},%
},%%
]
table[header=true, x index=1, y index=2, row sep=\\] {
Nr x y  \\
1 1 3             \\%0
2 0 2              \\%1
3 2 2              \\%2
4 2 1              \\%3
};

% Tabellarische Zeichnung von Annotationen =================

\addplot[no marks,
nodes near coords={},
%
% Versuch ......
visualization depends on={value \thisrowno{0} \as \punktnummer},
nodes near coords style={minimum size=15cm,overlay,
path picture={%\pgftransformreset % mir ist nicht klar, ob das hilft
\pgfmathtruncatemacro{\X}{\punktnummer-1}
\pgfmathtruncatemacro{\Y}{\punktnummer+1}
\node[anchor=north,red] at (P\Y){\punktnummer};
\path let \p1=($(P\Y)-(P\punktnummer)$),\p2=($(P\X)-(P\punktnummer)$),
\n1={atan2(\y1,\x1},\n2={atan2(\y2,\x2}
in [get angle path=\mytest];
\ifnum\mytest=0
\draw pic [draw,-latex,red,angle radius=1cm,fill=blue!50] 
{angle = P\Y--P\punktnummer--P\X};
\else
\draw pic [draw,-latex,red,angle radius=1cm,fill=blue!50] 
{angle = P\X--P\punktnummer--P\Y};
\fi
}},
]
table[header=true, x expr =0, y expr=0]{
Nr.
1
2
};

\end{axis}

% Test 
\draw[brown] (P0) circle[radius=5mm] node[right]{geht};
\draw[brown] (P1) circle[radius=5mm] node[right]{geht};
\draw[brown] (P1) -- (P3) node[midway, right]{geht auch};

\end{tikzpicture}
\end{document}