Random Extract duplicates features

[pinilla66]

What is the bug or the crash?

I've got a vector layer with 2705 features from which I want to randonmly extract 1000 of them. Using the Random Extract tool, it returns a set of 1000 features but some of them are duplicated (up to four times). I've checked the id attribute and it is correctly set.
For being on the safe side, I've also used the Random selection tool to verify if it returns a correct set of features and it has worked well.

Steps to reproduce the issue

1. Go to Toolbox
2. Click on Random extract
3. Select vector layer
4. Set 1000 features

Versions

Versión de QGIS
3.22.16-Białowieża
Revisión del código de QGIS
6f08e4d
Versión Qt
5.15.3
Versión de Python
3.9.5
Versión de GDAL/OGR
3.6.2
Versión de PROJ
9.1.1
Versión del registro de base de datos EPSG
v10.076 (2022-08-31)
Versión GEOS
3.11.1-CAPI-1.17.1
Versión de SQLite
3.39.4
Versión de PDAL
2.4.3
Versión del cliente de PostgreSQL
14.3
Versión de SpatiaLite
5.0.1
Versión de QWT
6.1.6
Versión de QScintilla2
2.13.1
Versión del SO
Windows 10 Version 2009

Complementos activos de Python
active_fire
0.3
DataPlotly
3.9.2
DEMto3D
3.51
HCMGIS
23.2.1
latlontools
3.6.7
OSMDownloader
1.0.3
pg_raster_import
3.1.0
pointsamplingtool
0.5.4
ProjectPackager
0.5.1
qdraw
3.0.2
qgisnetworklogger
0.2.0
QuickOSM
2.1.1
quick_map_services
0.19.32
sigpac_downloader
0.3
db_manager
0.1.20
grassprovider
2.12.99
MetaSearch
0.3.5
processing
2.12.99
sagaprovider
2.12.99

Supported QGIS version

• I'm running a supported QGIS version according to the roadmap.

New profile

• I tried with a new QGIS profile

Additional context

No response

[roya0045]

The algorythm clearly wants to select duplicates, I doubt this is a bug in any way.

[pinilla66]

Well, I can't imagine why anyone would want to extract the same feature twice or more times for geographical analyses, but it's probably my ignorance. Should I close this thread, then?

[roya0045]

I wager that this ticket could be converted in a Feature Request to have an option to not generate duplicates.

In the meantime the select & export option can't generate duplicates.

As far as why duplication is allowed, I'm not sure.

[agiudiceandrea]

@roya0045 anyway it seems to me there is an inconsistency: both the "Random extract" and the "Random selection" algorithms have the same description in the short help and in the documentation, but they behave differently.

Moreover the "Random extract" doesn't respect the number of feature to extract, even counting the duplicates.

Trying with a simple layer with 10 features testpoints.zip and trying to extract e.g. 7 features, the resulting layer will not always contain 7 features, but it could also contain less then 7 features.

Moreover also the "Random extract within subsets" algorithm behaves differently from the "Random extract" alg: it seems it extracts always the exact requested number of features without duplicates like the "Random selection" and the "Random selection within subsets" algorithms do.

So it seems we have three algs ("Random selection", "Random selection within subsets" and "Random extract within subsets") that behaves in a way, and only one algorithm ("Random extract") that behaves in a different way...

Maybe @alexbruy could shed some light on this.