pandoc documentation says:

A header without an explicitly specified identifier will be automatically assigned a unique identifier based on the header text.

However, pandoc assigns an identifier even if identifier is already set, causing incorrect HTML output:

$ ~/.cabal/bin/pandoc --version
pandoc 1.15.0.6
Compiled with texmath 0.8.3, highlighting-kate 0.6.
...

$ cat input.md 
# Chapter 1 {id="ch01"}

text

$ ~/.cabal/bin/pandoc -f markdown -t html -o output.html input.md 

$ cat output.html 
<h1 id="chapter-1" id="ch01">Chapter 1</h1>
<p>text</p>

Note: h1 tag has two id attributes.

@van-de-bugger,

from this source:

# Chapter 1 {#ch1}
# Chapter 2 {id="ch2"}

you will have this HTML output (using latest stable pandoc):

<h1 id="ch1">Chapter 1</h1>
<h1 id="chapter-2" id="ch2">Chapter 2</h1>

Extended Markdown requires identifiers tagged as {#identifier} and classes as {.class}.

Just in case it might help.

Thanks, I am aware about {#id} syntax, but I think {id="id"} is also valid.