Bug report #22038

Inconsistent intervals for similar labels in graduated classification

Added by Mic Del 6 months ago. Updated 6 months ago.

Status:Open
Priority:Normal
Assignee:-
Category:Symbology
Affected QGIS version:3.4.6 Regression?:No
Operating System: Easy fix?:No
Pull Request or Patch supplied:No Resolution:
Crashes QGIS or corrupts data:No Copied to github as #:29852

Description

Hi

There is currently an inconsistent meaning behind a similar appearance when we make a graduated classification.

In the
legend
label   ... meaning
we see  for QGIS
------  -----------
a - b    a ≤ x ≤ b
b - c    b < x ≤ c
c - d    c < x ≤ d

Note that the first class has left inclusion but not the others.
We could wrongly expect from reading only the label :

a - b equals to a ≤ x < b
b - c equals to b ≤ x < c
c - d equals to c ≤ x < d

... or any other scheme actually.

So the same "class interval label" have different interpretations if they are the first or not.

We need a label notation where the endpoints inclusions are explicit, for example :

common   French
style    style
-------  -------
[a, b]   [a, b]
(b, c]   ]b, c]
(c, d]   ]c, d]

... or at least if we keep the basic "a - b" notation then we must use the same interval scheme for all classes.

Note also that some softwares (R for example) use left-open/right-closed intervals by default :

     label
--------------
                 meaning
------  ------  ---------
(a, b]  ]a, b]  a < x ≤ b
(b, c]  ]b, c]  b < x ≤ c
(c, d]  ]c, d]  c < x ≤ d

... and some others (like openJUMP) are left-closed/right-open :

     label
--------------
                 meaning
------  ------  ---------
[a, b)  [a, b[  a ≤ x < b
[b, c)  [b, c[  b ≤ x < c
[c, d)  [c, d[  c ≤ x < d

So we need :
  • to use a consistent scheme for all classes
  • to add an option to be able to choose the left-open or right-open scheme
  • to add an option to generate the label with the common notation, the French (Bourbaki) notation or any other notation.

Thanks

History

#1 Updated by Mic Del 6 months ago

See also "Data class groupings" #16983

Also available in: Atom PDF