QGIS3.4 Organize columns, data integrity compromised

[qgib]

Author Name: Soares SOares (Soares SOares)
Original Redmine Issue: 20673
Affected QGIS version: 3.5(master)
Redmine category:attribute_table
Assignee: Julien Cabieces

---

When you use 'Organize columns' and 'move selection to top' in atribute table you can crash your data integrity.

To reproduce this bug you will need a layer with more than 14 columns and more than
45 rows. That is the specification of our smaler table. We have tried with
a smaller table and we don't get any error.

1 - pick up the atribute table (postgres with geometry or shapefile),
right click in a column name and select 'Organize columns...'
2 - Reorder the columns. Thinking in column position like a list, try a
sequence that bring to initial positions the end elements. In a example
14,12,5,1,2...
3 - Back to the table, start edition.
4 - Click at QGIS numerator of feature in table. this will turn editable the cell of
the column 14 (remind the example above),
5 - Chose in menu table 'move selection to top'
6 - look for the atribute's of column 14, it will apear in another column
of the same selected object.

---

• sample_data.zip (Soares SOares)

---

Related issue(s): #28553 (relates)
Redmine related issue(s): 20733

---

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

Does it happens on 2.18?

---

• status_id was changed from Open to Feedback

[qgib]

Author Name: Soares SOares (Soares SOares)

---

Yes, Giovanni. And also in 3.4LTR

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

Soares SOares wrote:

Yes, Giovanni. And also in 3.4LTR

my question was about 2.18, not 3.4

[qgib]

Author Name: Soares SOares (Soares SOares)

---

Yes, it happens in 2.18.

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

Soares SOares wrote:

Yes, it happens in 2.18.

I assume that the raw data integrity is not compromised(?)

---

• priority_id was changed from High to Normal
• crashes_corrupts_data was changed from 1 to 0

[qgib]

Author Name: Alessandro Pasotti (@elpaso)

---

I cannot reproduce on current master, can you please add a sample project and sample data and a video that demonstrate the issue?

[qgib]

Author Name: Soares SOares (Soares SOares)

---

Ok, I will send to you on monday.

[qgib]

Author Name: Soares SOares (Soares SOares)

---

Hello Alessandro. At the moment I can't make a video, but there is a project with a shapefile.
The columns already have been organized, so after open the project:

• Open attribute table;
• Start edition;
• Select any feature (click on QGIS's numerator in table);
• Move selection to top.

The content atribute in selected cell (col_18) will be sended to 'col_11'

---

• 13894 was configured as sample_data.zip

---

• sample_data.zip (Soares SOares)

[qgib]

Author Name: Soares SOares (Soares SOares)

---

2.14 just fine

[qgib]

Author Name: Alessandro Pasotti (@elpaso)

---

Confirmed on master thanks for the demo data.

---

• version was changed from 3.4.2 to 3.5(master)

[qgib]

Author Name: Alessandro Pasotti (@elpaso)

---

• relates was configured as Attribute table: Order of columns doesn't persist #28553

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

• status_id was changed from Feedback to Open

[qgib]

Author Name: Luigi Pirelli (@luipir)

---

is it confirmed data loss or compromising? if so, the issue shouldn't have normal priority

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

• priority_id was changed from Normal to High
• crashes_corrupts_data was changed from 0 to 1

[qgib]

Author Name: Julien Cabieces (Julien Cabieces)

---

Proposed PR : #8725

[qgib]

Author Name: Luigi Pirelli (@luipir)

---

please Soares SOares give a double check to the proposed PR if it's possible. otherwise you have to wait merge and nightly build

Julien Cabieces wrote:

Proposed PR : #8725

[qgib]

Author Name: Julien Cabieces (Julien Cabieces)

---

Applied in changeset fba4b25.

---

• status_id was changed from Open to Closed
• done_ratio was changed from 0 to 100

[qgib]

Author Name: Soares SOares (Soares SOares)

---

I've Tested 2.18.27 and the bug still there. The bug also persists in 3.4.3 I've used the same file in bug report.

---

• status_id was changed from Closed to Reopened
• assigned_to_id was configured as Julien Cabieces

[qgib]

Author Name: Julien Cabieces (Julien Cabieces)

---

The correction had not been backported in 2.18.27 but is present in 3.4.4 (not 3.4.3)

---

• status_id was changed from Reopened to Closed
• resolution was changed from to fixed/implemented