Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Rapidxml namespace strip patch #295
Applies the patch mentioned on sourceforge for rapidxml to nuke the namespace prefixes for element names and attributes. As usual, this arises in creative spreadsheets written by 3rd party tools. However, Excel and the Apache POI-using R packages can read these files and, therefore, so should readxl.
The other solution would be to use a rapidxml fork that handles namespaces properly. That seems like a much bigger change and, more importantly, would require C++11. So far readxl has done without that, so I assume we should continue?
I've probably used the new flag in more places than is strictly necessary, but it's hard to anticipate where the novel prefixes will show up. I'm still trying to get more problematic specimens for testing, but it's hard. It's usually niche BI tools that do this and people aren't able to share the sheets.
I reached out to the author of the patch, to say "thanks" at the very least.