Datatool (datagidx) index with special characters

Question

When I define multiple terms that contain an underscore and begin with the same string ex_ 1, ex_ 2 (as shown in the example below) I get the following error:

! TeX capacity exceeded, sorry [input stack size=5000]. \expandafter l.30 \end{frame}

However, when I define the term e2 using the commented out line below (e2_ 2) it compiles fine. If I just remove the escaped underscores (see below) that works as well. Can someone tell me how to work around this?

\documentclass[16pt,pdftex]{beamer}
\usepackage[T1]{fontenc}

\makeatletter
% beamer doesn't define \@idxitem, but it's used by the
% \printterms "index" style, so define it:
\newcommand\@idxitem{\par\hangindent40pt}
\makeatother

\usepackage{datagidx}

\newgidx{index}{Index}
\DTLgidxSetDefaultDB{index}

% this fails:
 \newterm[label=e1]{ex\_ 1}
 \newterm[label=e2]{ex\_ 2}

% this works
%\newterm[label=e1]{ex\_ 1}
%\newterm[label=e2]{e2\_ 2}

% this works
%\newterm[label=e1]{ex 1}
%\newterm[label=e2]{ex 2}
\begin{document}

\begin{frame}
\frametitle{example}
\gls{e1}
\gls{e2}
\end{frame}

\begin{frame}[allowframebreaks]
\frametitle{Index}
\printterms[style=index]
\end{frame}
\end{document}

Answer 1

The second comment by Nicola Talbot answered my question: Without specifying a sort option it tries to sort by the term and when the sorter gets to the escape character it has a hard time. Specifying the sort option:

\newterm[label=e1,sort={ex1}]{ex\_ 1}
\newterm[label=e2,sort={ex2}]{ex\_ 2}

Solves the problem for me.