[Tex/LaTex] PGFplots axis ticks numbers format

Question

I would like to know how could I exclude 0 of this axis format

yticklabel style={/pgf/number format/.cd,sci,sci zerofill,precision=1}

In order to avoid this.

the most similar what I've got is using this code (as MWE)

\documentclass[8pt,a4paper,dvipsnames]{article}
\usepackage[utf8]{inputenc}
\usepackage[left=2cm,right=2cm,top=1.5cm,bottom=1.5cm]{geometry}

\usepackage{xcolor}



\usepackage{pgfplots}
\pgfplotsset{compat=1.14}

\begin{document}
\begin{figure}
\centering
\begin{tikzpicture}
\begin{axis}[grid=major, scaled ticks=false,scale=2.2, y post scale=2, yticklabel style={/pgf/number format/.cd, 1000 sep={\;}}]
\addplot+[
        only marks,
        mark=+,
        color=NavyBlue,
        error bars/.cd,
        x dir=both, x explicit,
        y dir=both, y explicit,
        ]
        table[y error=yerror,x error=xerror]
            {NmP3pgf1.dat};
\addplot[
        red,
        domain=-3:1
        ] {-120322.13*x + 53246.18};  
\end{axis}
\end{tikzpicture}
\caption{pgfplots}
\end{figure}
\end{document}

Which produces this

being this NmP3pgf1.dat

x           y                xerror       yerror  label
0.00        52080.65         0.01         648.31   a
-0.80       148381.84        0.01       2618.50   a
-2.80       391752.88        0.01       9554.95   a

Answer 1

Essentially Jake's solution is correct as usual but here your numbers are too big for the \ifdim test. Hence a bit more lower level pgf number formatting is needed.

If you just use the following it will perform a test for zero and execute the correct branch of the if.

yticklabel={\pgfkeys{/pgf/fpu}%
    \pgfmathparse{\tick}%
    \pgfmathfloatifflags{\pgfmathresult}{0}{0}{%
        \pgfmathprintnumber[sci,sci zerofill,precision=1]{\tick}%
    }%
    \pgfkeys{/pgf/fpu=false}%
}