You may try with the following code:
\documentclass{article}
\usepackage{mathtools}
\newcommand{\Rint}{\mathop{\mathrlap{\pushR}}\!\int}
\newcommand{\pushR}{\mathchoice
{\mkern2.5mu R}
{\scriptstyle R}
{\scriptscriptstyle R}
{\scriptscriptstyle R}
}
\begin{document}
\[
\left|\Rint_a^b f\right|\textstyle\left|\Rint_a^b f\right|_{\left|\Rint_a^b f\right|}
\]
\end{document}

The absolute value is only for demonstration purpose.
The main problem here is to adapt the R to the various formula styles, so we use \mathchoice
. With \mathrlap
we can build a zero width box that will not contribute to the width of the formula. Since \int
is an operator, we put this zero width formula into a \mathop
, then \!
is used to remove the space that TeX inserts between two consecutive Operator atoms.
Next we define \pushR
, that's simply a space in display style, followed by a normal size R. In text and lower styles we simply put a reduced size R. Probably the result will not be nice in \scriptscriptstyle
, but it doesn't seem so important.
The same idea can be used for the principal value integral:
\newcommand{\pvint}{\mathop{\mathrlap{\pushpv}}\!\int}
\newcommand{\pushpv}{\mathchoice
{\mkern5mu\rule[.6ex]{.5em}{1pt}}
{\mkern2.8mu\rule[.5ex]{.35em}{.8pt}}
{\mkern2.5mu\rule[.29ex]{.3em}{.7pt}}
{\mkern2mu\rule[.2ex]{.2em}{.5pt}}
}
