Virtual fields: graduated symbology with natural breaks does not use correct limits

[qgib]

Author Name: Giovanni Manghi (@gioman)
Original Redmine Issue: 10995
Affected QGIS version: master
Redmine category:virtual_fields

---

Steps:

• pick a polygon layer
• add a virtual field for area (two decimal places), like $area/10000 to get it in HA
• clone the resulting column in a "real" column

now test the virtual field and the real column to make a graduated symbology using natural breaks: the ranges for the "real" field make sense while the ones for the virtual field seems wrong.

---

Related issue(s): #19342 (relates)
Redmine related issue(s): 11000

---

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

NOTE:
maybe because this issue, maybe because it seems that when doing symbology the decimals of the virtual field are ignored (it seems to use always the maximum number of decimals regardless the choice that was made when creating the field), I'm noticing that when doing graduated symbology based on VF (testing polygons and $area function) then sometimes there are features that do not render... but if you toggle editing then they render...

[qgib]

Author Name: Matthias Kuhn (@m-kuhn)

---

I do not understand the bug report, but I think it is related to #19342 like #19336.

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

Matthias Kuhn wrote:

I do not understand the bug report

• pick a polygon layer and compute the $area of polygons in a field and a virtual field with 2 decimals

• then do the graduated symbology using those fields, with natural breaks option

• with the "real" field the classes are like

221832.3600 - 280204.1300
280204.1300 - 430749.4500
430749.4500 - 553529.0200

the graduated renderer always shows 4 decimals, but anyway the values are correct -> the ones in the "real" column

• with the Virtual field the classes (natural breaks) are like

221832.3558 - 280204.1329
280204.1329 - 430749.4541
430749.4541 - 553529.0204

that is not correct because those are not the values that are in the column used for the symbology.

[qgib]

Author Name: Matthias Kuhn (@m-kuhn)

---

Thank you for the clarification.

So it is indeed the same problem that the precision is not taken into account when evaluating the expression. If #19342 is solved, this issue should neither show up any longer.

[qgib]

Author Name: Paolo Cavallini (@pcav)

---

• subject was changed from Virtual fields: graudated symbology with natural breaks does not use correct limits to Virtual fields: graduated symbology with natural breaks does not use correct limits

[qgib]

Author Name: Matthias Kuhn (@m-kuhn)

---

Hi Giovanni,

Could you test #1535 to see if it fixes this problem?

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

Matthias Kuhn wrote:

Hi Giovanni,

Could you test #1535 to see if it fixes this problem?

Hi Matthias, thanks for the patch!

I will not be able to test it in the next weeks... I will give it a try as soon as I can, if anyone hasn't looked at it before. Cheers!

[qgib]

Author Name: Matthias Kuhn (@m-kuhn)

---

Fixed in changeset "7318c73012a2c7245f5ac7a268a457ae84f0a4de".

---

• status_id was changed from Open to Closed