Closed palmerj closed 1 week ago
Hi @palmerj,
can you try with newer version of QGIS? Only thing weird I noticed is that you added duplicated indexes with different names at the end of your SQL.
CREATE INDEX fki_fkey_parcel ON title_parcel_association USING btree (par_id);
CREATE INDEX fki_fkey_title ON title_parcel_association USING btree (ttl_title_no);
CREATE INDEX idx_tpa_par_id ON title_parcel_association USING btree (par_id);
CREATE INDEX idx_tpa_ttl_title_no ON title_parcel_association USING btree (ttl_title_no);
Hi @phidrho
can you try with newer version of QGIS?
I've updated to 3.28.0
Only thing weird I noticed is that you added duplicated indexes with different names at the end of your SQL.
AH good find. I've removed those.
However, still the same result.
@palmerj
Can you also please check the new GDAL version? There's been an upgrade to 3.6.0 recently which maybe still isn't available on Mac, but will sure be in next days/weeks.
I didn't think that removing duplicated indexes would fix this behavior, because it's probably not related. Can you send some dump of data? Few parcels and few related titles would be enough, so I'll try to replicate it on Windows. If your data is sensitive please make up some fake data that fits your schema so that I and developers can test it properly on our side.
EDIT: Missing Relation in configuration - I think that this message also isn't related to problem, as I also get this message when using M-to-M relation in GPKG, but everything works as expected.
Ok great. I will prepare the data tomorrow.
Note the QGIS app bundle version of GDAL is 3.3. See:
QGIS version | 3.28.0-Firenze | QGIS code revision | ed3ad0430f |
---|---|---|---|
Qt version | 5.15.2 | ||
Python version | 3.9.5 | ||
GDAL/OGR version | 3.3.2 | ||
PROJ version | 8.1.1 | ||
EPSG Registry database version | v10.028 (2021-07-07) | ||
GEOS version | 3.9.1-CAPI-1.14.2 | ||
SQLite version | 3.35.2 | ||
PDAL version | 2.3.0 | ||
PostgreSQL client version | unknown | ||
SpatiaLite version | 5.0.1 | ||
QWT version | 6.1.6 | ||
QScintilla2 version | 2.11.5 | ||
OS version | macOS 12.6 | ||
Active Python plugins contour | 2.0.10 QGIS3-getWKT | 1.4 CityJSON-loader | 0.8.1 QuickWKT | 3.1 searchlayers | 3.0.12 qgis-maptiler-plugin | 3.1.1 nominatim_locator_filter | 0.2.4 processing | 2.12.99 sagaprovider | 2.12.99 grassprovider | 2.12.99 db_manager | 0.1.20 MetaSearch | 0.3.6
I can confirm this isn't yet fixed -- the code has not changed at all in recent versions
Thanks @nyalldawson.
@phidrho do you still want the data to support development work?
@palmerj I am not currently a developer, but it's always good to give an example data to developers in bug reports so that they can test it and recreate behavior on their machine, which leads to much faster determination of bug in code.
Small PostGIS test dataset here linz_test_data.sql.zip. Licence terms in the header of the SQL file.
Note while it would be good to get the 3rd join, we actually have relations to plain tables as many as 8 joins. This is due to highly normalised structure of the database.
Hi @palmerj,
I finally got some time to check your data, and everything seems OK with data, but everything also seems to work as documented.
Unfortunately, I didn't manage to find where in manual is noted that you can recursively loop through nested features:
The manual seems to show it's possible with https://docs.qgis.org/3.22/en/docs/user_manual/working_with_vector/attribute_table.html#introducing-many-to-many-n-m-relations.
You can set up each layer's attribute form to show you data from mapped M-to-M feature, but not with Identify Results Panel
.
So, this issue seems to be a feature request
and not a bug report
.
In the meantime, you can use Actions for relations plugin which can fulfill your needs, but first you need to add some more relations manually which connect a mapping table
to data tables
- I called them fkey_map_parcel
and fkey_map_title
- see picture:
I recorded you an example how to use plugin.
Ok, maybe it's not clearly documented that the feature info supports n-m-relations, so let's call it a feature!
Ok, maybe it's not clearly documented that the feature info supports n-m-relations, so let's call it a feature!
I agree, it's kind of same from user perspective view, but different from QGIS organization view.
Don't get me wrong, I really like the idea, and maybe your institution/organization has budget to invest into this feature.
Here is a list of companies that offer commercial support for implementation of new features.
What is the bug or the crash?
The manual seems to show it's possible with https://docs.qgis.org/3.22/en/docs/user_manual/working_with_vector/attribute_table.html#introducing-many-to-many-n-m-relations.
However, when using the identify tool, I can't get the table attributes for the 3rd table to show using the join table. See
In the example above I should be able to expand the relationship using field on ttl_title_no field, but it's not an option.
Not sure if it's related but I see the following in message log (repeated each time I click on a feature):
The relationships seem to be formed ok.
Steps to reproduce the issue
Versions
Active Python plugins QGIS3-getWKT | 1.4 CityJSON-loader | 0.6.2 QuickWKT | 3.1 qgis-maptiler-plugin | 1.1.4 nominatim_locator_filter | 0.2.3 processing | 2.12.99 sagaprovider | 2.12.99 grassprovider | 2.12.99 db_manager | 0.1.20 MetaSearch | 0.3.5
Supported QGIS version
New profile
Additional context
No response