Inclusion of hash marks in a LaTeX \indexentry aliased with a \newcommand

This is a strange one. I've had to do some interesting workarounds to get it to function even a little bit, but we have another stumbling block.

I've been struggling to create a separate document which employs a large .ind file (index created using makeindex from catenated, small individual .idx files), the idea being that I will eventually have a single document which indicates the SUB-document开发者_JS百科s and page numbers where the indexed data occurs.

To this end, I've had to define a command which collects all index entries as:

\newcommand{\myindexer}[3]
    {\index{myindex}{#1 : #2.#3!\href{\doctitle.pdf}}}

The result is index entries in the myindex.idx file as:

\indexentry{IndexedItemA : 55.iii!\href{Volume 1.pdf}{Volume 1.pdf}}{30}}
\indexentry{IndexedItemB : 23.vi!\href{Volume 3.pdf}{Volume 3.pdf}}{114}}

Yet, the problem exists in trying to create a hyperref to the target PDF. In order to open a PDF at a specific page, the following format needs to be employed (thanks to someone else on this board for pointing out the #page.xx syntax in the hyperref package):

\href{FILENAME#page.XX}{Link Text}

wihch means that a hash mark needs to be included in the output stream of the \newcommand when myindex is declared, possibly using an escaped hash mark # as:

\newcommand{\myindexer}[3]
    {\index{myindex}{#1 : #2.#3!\href{\doctitle.pdf\#page.\thepage}}}

or according to some fragment I found somewhere using a double-hash mark as:

\newcommand{\myindexer}[3]
    {\index{myindex}{#1 : #2.#3!\href{\doctitle.pdf##page.\thepage}}}

The former results in the following error during compilation of the latex source for the multiple documents being indexed:

! Illegal parameter number in definition of \@gtempa.
<to be read again>
                   \thepage
l.31 \myindex{IndexedItemA}{55}{iii}

While the latter results in an unwanted backslash in the resulting myindex.idx file as:

\indexentry{IndexedItemA : 55.iii!\href  {Volume 1.pdf\#page.33}{Volume 1}}{33}

which results in the target PDF not opening correctly.

Any idea how the newcommand can be forced to output the hash mark to support hyperlinking of PDFs in this way?

What about wrapping the \href in \protect?

You can swap the catcode of # and %:

\catcode`\%=6 \catcode`\#=12

Then use % to mark parameters:

\newcommand{\myindexer}[3]
    {\index{myindex}{%1 : %2.%3!\href{\doctitle.pdf#page.\thepage}}}

After that, restore the usual catcodes:

\catcode`\%=12 \catcode`\#=6

This is kind of clumsy, but it should work even if something interferes with usual escaping mechanisms, and it allows \href to be exanded at the point \myindexer is used.