Search code examples
javaxmlncbipubmed

Is Pubmed returning invalid XML results?

I am using JEUtils to fetch and parse Pubmed results in Java (it's a tool which seems to be abandoned).

Since a few days ago the tool is throwing exceptions in some results, and upon inspection it seems that Pubmed is not respecting its own DTD (the first one in its DTDs Page).

For example, this article, whose XML result you can see by Format: XML or see directly here has the following elements:

 <DataBankList CompleteYN="Y" />
 <GrantList CompleteYN="Y" />

But according to the DTD these elements must have at least one nested element:

<!ELEMENT   DataBankList (DataBank+)>
<!ELEMENT   GrantList (Grant+)>

Question: am I missing something, or is Pubmed creating result XMLs which do not comply with their own DTD?

Solution

  • Looks to me like you're right. The 2016-01-01 NLMMedlineCitationSet DTD seems to be the most recent version (see also https://www.nlm.nih.gov/databases/dtd/history_dtd_nlmmedline.html), and it does seem to demand one or more child elements for these two elements.

    I would use the "Contact NLM" link at https://www.nlm.nih.gov/bsd/licensee/elements_descriptions.html to open a support ticket.