I think this is the one you are looking for:
http://yonik.com/solr/atomic-updates/
On Wed, Dec 4, 2019 at 9:40 AM Nan Yu <n...@z-geoinfo.com> wrote:
> Hi, I'm new to solr and my question might seems to be dumb. Please
> bear with me.
> In short, if a file (after indexed by solr) changed its name, could I
> only update resourcename and keep all other information the same in the
> index? Or do I have to delete it and then re-index the file?
> Below are the detailed version of my question: My problem: I have
> many (tens of thousands) files that are sort of messed up with file
> name/location: (1) wrong file type in file name: for example, some files
> are pdf file but are named as .txt. (2) wrong directory of file: a file
> named 001.pdf should be under a specific folder, such as: abc/001.pdf. The
> "abc" could be found in 001.pdf. But now, this file might be placed under
> xyz/001.pdf. So, I would like to relocate the file from xyz/001.pdf to
> abc/001.pdf.
> I need to correct those errors and also being able to search the
> content of those documents.
> How I was trying to solve the issue using solr:(1) index all files(2)
> from content_type and/or application_name, I could know the correct file
> type (pdf, text etc.) and compare it with the resourcename. Found out those
> files that I need to update.(3) query all "abc", "xyz" etc. and find out
> those files that need to be relocated.(4) rename/relocate files from step
> (2) and (3)(5) update solr index.
> My question:I tried to use the solr UI to update resourcename. I was
> able to do so but the problem is the id now has a new version that only
> contains resourcename.
> Here is how I made the update:
> In the "Documents" page: Request-Handler: /updateDocument Type:
> JSON:Document(s): {"id":"/testproject/abc/001.txt",
> "resourcename":{"set":["/testproject/abc/001.pdf"]}}
>
> After submission, and query/browse the file, the only thing left
> is resourcename. All other information, such as content_type, created etc.
> are gone. If I could not update the " resourcename" (and keep all
> other information), it means that I either run a script to delete the
> document that I renamed/relocated and re-index them, meaning that myscript
> will do: find the file, rename/relocate, delete the original index,
> re-index/post the new file. Is this the best approach?
> Thanks!Nan
