Hi!

I just ran into the same problem and solved it somewhat manually, but it worked well for a 200+ pages document.

Basic approach: unzip the docx file (these are zip-archives with a different extension to the filename), then tweak it a bit to put a text entry of the index we can later use in the file converted by Pandoc, re-zip the beast and the use Pandoc to create a LaTeX file. With the below scripts, the converted file would have a '++index{Index entry}' for each entry. Of course, the '++index' needs to be find-and-replaced to '\index'. Done.

Here the details (please excuse the Markdown, I copied from my personal wiki). Hope this helps the one or other out there…

# Preserving indices
Pandoc does not do indices. So to keep them, unzip the Word file, open `document.xml` in Atom, and replace the index entries with the LaTeX command.

First make all xml-tags stand on one line. And replace all index entries by a LaTeX-command. I am using `++` instead of the `\` to make later replacement in the TeX-file easier. _Of course, check after conversion in Pandoc that the `++` needs to me manually turned into a `\` for the index-commands to work._

Now the difficult part: any lines before and after a line starting with `++index` need to be removed, from and until a line starting with something else than `<`. And, there could be several index commands running into each other without any normal text between.

So we pull out any index entry and make a `++index` out of it using `sed`, which is dirty, but quick. Then perl is dropped onto the file to pull the `++index` before any xml-tags, which puts it right behind the normal text, where they should go later. Then we write the rest out as is, to make sure all xml-tags Word kept open are properly closed. One small alteration: any text in the Word index command is simply replaced by _FOO_, so it would be easier to track if anything went wrong or these indices somehow pop up again.

This can be achieved with the following perl-script saved in WordIndex2LaTeX.pl (and made executable, so chmod +x WordIndex2LaTeX.pl):
```
#!/usr/bin/perl

# @Author: Hendrik G. Seliger
# @Date:   4 February 2022, 11:34 +01:00
# @Filename: WordIndex2LaTeX.pl
# @Last modified time: 4 February 2022, 11:36 +01:00
# @License: GPL3
# @Copyright: © Copyright 2022 by Hendrik G. Seliger

###

$keptlines='';
$indexlines='';

while ( <STDIN> ) {
        if ( ( $_ =~ /^</ ) || ($_ =~ /^XE / ) ) { # line with xml tag or word index entry
                $keptlines .= $_; # save the line
        } elsif ( $_ =~ /^++index/ ) {
                # Found an index entry. Now, put the LaTeX command BEFORE the
                # Word tag, so that the tags are correctly opened and closed, but
                # the LaTeX command appears first
                $indexlines .= $_; # save the line
        } else { # normal text line, print all kept ones and current, erase memory
                print $indexlines;
                print $keptlines;
                print $_;
                $keptlines='';
                $indexlines='';
        }
}
print $indexlines;
print $keptlines;
```

So hence the conversion is achieved with

```
mkdir D
cd D
unzip ../MyDoc.docx
cd word
cat document.xml|sed -E 's/>/>\n/g' |sed -E 's/(.)</\1\n</g' | sed -E 's/^XE &quot;(.*)&quot;/++index{\1}\nXE \&quot;FOO\&quot;/g' | ../../WordIndex2LaTeX.pl| tr -d '\n' >document2.xml
```

Back up `document.xml` and rename `document2.xml` to `document.xml`. Re-zip the document
```
mv document.xml ../..
mv document2.xml document.xml
cd ..
zip -r ../D.docx *
cd ..
```