load.xml can't load entity reference. about apoc HOT 4 OPEN

@vunhatchuong123 Thanks for reporting. We will investigate and come back to you.

from apoc.

@vunhatchuong123 First a caveat; I'm part of the team working with APOC but I have not very much experience with XML in particular. I wonder if this is really a valid XML file. I tried to upload your XML data to two different online XML formatters, https://www.freeformatter.com/xml-formatter.html and https://jsonformatter.org/xml-formatter, to see how they behave compared to APOC. The first one errors with the similar error Unable to parse any XML input. Error on line 3: The entity "Ouml" was referenced, but not declared.. The second one does accept &Ouml but render it as is rather than format it into an Ö. Do you have earlier experience where XML handling works as you expect it?

Best regards Louise Söderström

Ps. I do see the usefulness of your request, myself having 2 Ö in my name. ;)

from apoc.

I stopped using neo4j right now so I won't be able to help that much but I'll try my best.

I think the problem comes from APOC not able to process dtd type definition files, specifically in this case it's dtd entity definitions. This dataset comes from DBLP, and in it there's a dblp.dtd file.

Here's a preview of that file:

<!ENTITY Ouml    "&#214;" ><!-- capital O, dieresis or umlaut mark -->
<!ENTITY Oslash  "&#216;" ><!-- capital O, slash -->
<!ENTITY Ugrave  "&#217;" ><!-- capital U, grave accent -->
<!ENTITY Uacute  "&#218;" ><!-- capital U, acute accent -->

So because APOC fails to load dblp.dtd, it doesn't understand &Ouml.

from apoc.

Thanks for coming back with more information. We have made a conscious decision not to support DTD files of security reasons. I will see if we can improve our documentation around this so it is more clear it is not supported.

from apoc.