DOIs not found when using query batch deposits

We are trying to link our articles’ references using query batch deposits. Unfortunately, this method seems to be unable to resolve very obvious citations that other methods easily find DOIs for.

For example, if we upload this using a doQueryUpload operation at (headers omitted for clarity):

<query key="rb11" enable-multiple-hits="true">
  <unstructured_citation>KIRBY, Douglas (2002) «The Impact of Schools and School Programs Upon Adolescent Sexual Behaviour», Journal of Sex Research, vol. 39, no 1, p. 27-33.

It gives this result:

      <query key="rb11" status="unresolved" fl_count="0">
        <article_title>«The Impact of Schools and School Programs Upon Adolescent Sexual Behaviour», Journal of Sex Research, vol</article_title>
        <journal_title>39 no 1</journal_title>

However, going the metadata deposit route including citations (doMDUpload), the metadata submission result contains the right DOI:

 <citation key="rb11" status="resolved_reference">10.1080/00224490209552116</citation>

Even stranger, the DOI is resolved correctly if I upload the exact same query batch as with doQueryUpload, but with the « Interactive query upload tool » at /servlet/submissionAdmin?sf=showQueries

<query key="rb11" status="resolved">
<doi type="journal_article">10.1080/00224490209552116</doi>

Please note I have tried different combinations with the enable-multiple-hits and secondary-query attributes and that they all gave the same result when uploading through the doQueryUpload method: DOIs are not resolved.

Also note that it doesn’t seem to be an isolated case, as initial tests show that most of our citations do not seem to resolve event if they have “obvious” DOIs.

Thanks for any help,
Béranger Enselme

Hi Béranger,

Thanks for your questions, and apologies for the delay. Crossref had its annual staff retreat recently, so we’re still catching up on some things we missed while we were away.

I will look into this further, but it’s very possible that the XML queries (whether done via a file upload or via HTTPs) use different underlying citation matching than the citation matching used when references are submitted as part of metadata deposits.

Can you tell me more about your use case for the ‘doQueryUpload’ XML queries?


Of course. We publish articles from hundreds of journals, but deposit DOI metadata only for a fraction of those.

Most journals register their DOIs themselves, or do not register DOIs at all. We still try to link the references of their articles to DOIs, and the only options Crossref offers for that is either doQueryUpload or, both of which find very few DOIs for references when compared to doMDUpload.

Thanks for your help!

Thanks for that additional context.

I see that we’ve already informed on product management team about this, and they’ll take it up with the technical team when time and resources allow.

You can follow any progress here


1 Like