Re: solr tika extraction video creation date problem (hours ahead)
Thank you very much Alex for the great suggestion. On Fri, Apr 5, 2019 at 7:25 PM Alexandre Rafalovitch wrote: > Well, Tika would use different libraries to extract different formats. > So maybe there is a bug. I would just get a standalone tika (of > matching version to the one in Solr) and see what the output from two > sample files are. Then, I would check with the latest Tika, just in > case. > > I would also use some non-Tika way to check what the dates are, just > in case the date is wrong during encoding rather than during indexing. > A low-probability chance, but just covering all the bases. > > Regards, >Alex. > > On Fri, 5 Apr 2019 at 01:39, whisere wrote: > > > > Thanks Alex. The problem is image creation date is correct, but the video > > creation date is wrong (hours behind), if I set the time_zone I think the > > image creation date will be wrong then. wonder what the difference > between > > image and video extraction in tika. > > > > > > > > -- > > Sent from: http://lucene.472066.n3.nabble.com/Solr-User-f472068.html >
Re: solr tika extraction video creation date problem (hours ahead)
Well, Tika would use different libraries to extract different formats. So maybe there is a bug. I would just get a standalone tika (of matching version to the one in Solr) and see what the output from two sample files are. Then, I would check with the latest Tika, just in case. I would also use some non-Tika way to check what the dates are, just in case the date is wrong during encoding rather than during indexing. A low-probability chance, but just covering all the bases. Regards, Alex. On Fri, 5 Apr 2019 at 01:39, whisere wrote: > > Thanks Alex. The problem is image creation date is correct, but the video > creation date is wrong (hours behind), if I set the time_zone I think the > image creation date will be wrong then. wonder what the difference between > image and video extraction in tika. > > > > -- > Sent from: http://lucene.472066.n3.nabble.com/Solr-User-f472068.html
Re: solr tika extraction video creation date problem (hours ahead)
Thanks Alex. The problem is image creation date is correct, but the video creation date is wrong (hours behind), if I set the time_zone I think the image creation date will be wrong then. wonder what the difference between image and video extraction in tika. -- Sent from: http://lucene.472066.n3.nabble.com/Solr-User-f472068.html
Re: solr tika extraction video creation date problem (hours ahead)
Sounds like timezone normalization issue. Possibly at Tika stage. Check what your SOLR_TIMEZONE variable set to. Not sure in which file. Regards, Alex On Thu, Apr 4, 2019, 12:50 AM Where is Where, wrote: > Hello , I was following the instruction > > https://lucene.apache.org/solr/guide/7_1/uploading-data-with-solr-cell-using-apache-tika.html > to upload files with metadata stored and indexed in solr. I was checking > the extracted creation date ( attr_meta_creation_date ), for image, jpg > etc, the creation dates are correct but all creation dates for video are 11 > hours ahead of the actual creation date. (The dates are correct when viewed > in other applications) It causes problem with searching due to this > inconsistency. Any idea is much appreciated. Thanks! >
solr tika extraction video creation date problem (hours ahead)
Hello , I was following the instruction https://lucene.apache.org/solr/guide/7_1/uploading-data-with-solr-cell-using-apache-tika.html to upload files with metadata stored and indexed in solr. I was checking the extracted creation date ( attr_meta_creation_date ), for image, jpg etc, the creation dates are correct but all creation dates for video are 11 hours ahead of the actual creation date. (The dates are correct when viewed in other applications) It causes problem with searching due to this inconsistency. Any idea is much appreciated. Thanks!