QGIS » QGIS Application

belg4mit - wrote:

Example files and results are available here https://nmrgroupinc-my.sharepoint.com/:u:/r/personal/jpierce_nmrgroupinc_com/Documents/Public/20033.zip?csf=1&e=XG2fDM (22MB)

it asks me to login in a Microsoft site, which I would rather not do. Do you have another link?

The sharing permissions were incorrect for some reason, try this:
https://nmrgroupinc-my.sharepoint.com/:u:/p/jpierce/ETsAu_igD9xFvPrWYG6myj8BG3S_GQ2UV-dApVZsoMOEjA?e=1YvA4P

belg4mit - wrote:

The sharing permissions were incorrect for some reason, try this:
https://nmrgroupinc-my.sharepoint.com/:u:/p/jpierce/ETsAu_igD9xFvPrWYG6myj8BG3S_GQ2UV-dApVZsoMOEjA?e=1YvA4P

everything looks normal here (also in your attached result), unless of course I'm not understanding what is wrong in results. But unless I'm blind (possibly) I'm no seeing it.

Did you check the screenshots I included? I neglected to spell it out,
but some of the fields are pretty obviously misaligned. For example, "NY"
(New York state) as the the ZIP Code, and the city of Ossining as the state.Similarly, there is no way anything has been constructed in (YR_BLT) 367400.

belg4mit - wrote:

Did you check the screenshots I included? I neglected to spell it out,
but some of the fields are pretty obviously misaligned. For example, "NY"
(New York state) as the the ZIP Code, and the city of Ossining as the state.Similarly, there is no way anything has been constructed in (YR_BLT) 367400.

I see that the wrong data (i.e. "NY" in the "mail_zip" field and large numbers in the "yr_blt" field) already in your INPUT layer (Ossining), so of course it yelds "wrong" results in the output of the join.

Hrmph, so they are. I could have sworn I checked after subsetting the original ogr2ogr output and it was fine. It seems any operation that generates a new file from these large geopackages, not just spatial join, causes the problem.

Here's the ogr2ogr output https://nmrgroupinc-my.sharepoint.com/:u:/p/jpierce/EcV_o8PsWalGgh-nuDKDRJYBd9ADzEHOHsiquZx8fkuQHg?e=3O8ytz which is properly aligned; I tried to use the Ossining subset to reduce the download and opening time (it's 810MB and almost three million records), but it has helped further frame the problem at least.

Here's the ogr2ogr output https://nmrgroupinc-my.sharepoint.com/:u:/p/jpierce/EcV_o8PsWalGgh-nuDKDRJYBd9ADzEHOHsiquZx8fkuQHg?e=3O8ytz which is properly aligned; I tried to use the Ossining subset to reduce the download and opening time (it's 810MB and almost three million records), but it has helped further frame the problem at least.

I downloaded this dataset, loaded it along with the administrative boundaries. On map I made a selection on both layers to define subsets, then in the "join by location" tool I enabled for both inputs the "selected features only" option. Used the spatial predicates you shown in the screenshot you attached here: results are correct.

Here's the ogr2ogr output https://nmrgroupinc-my.sharepoint.com/:u:/p/jpierce/EcV_o8PsWalGgh-nuDKDRJYBd9ADzEHOHsiquZx8fkuQHg?e=3O8ytz which is properly aligned; I tried to use the Ossining subset to reduce the download and opening time (it's 810MB and almost three million records), but it has helped further frame the problem at least.

redid the test with the full datasets (1846.00 seconds), results are correct here.

belg4mit - wrote:

I just tried with subsetting, and it still fails here. Are you also using Windows?

I'm on Linux

And generating a geopackage? Creating a geopackage from the two geopackages works, at least for a selected subset, when outputting to a shapefile.

I used as inputs this https://nmrgroupinc-my.sharepoint.com/:u:/p/jpierce/EcV_o8PsWalGgh-nuDKDRJYBd9ADzEHOHsiquZx8fkuQHg?e=3O8ytz and the boundaries you posted in a previous comment. The subset was done by selection (so I have not used QGIS or ogr2ogr to create physically a new sub-dataset).

Please test on QGIS master (I'm using QGIS master).

Hmm, after I was able to track down the installer, the process seems to work fine in the weekly build.