Bug report #21192
substr function in Select by Expression behaving strange
| Status: | Closed | ||
|---|---|---|---|
| Priority: | Normal | ||
| Assignee: | - | ||
| Category: | Expressions | ||
| Affected QGIS version: | 3.5(master) | Regression?: | No |
| Operating System: | Easy fix?: | No | |
| Pull Request or Patch supplied: | No | Resolution: | |
| Crashes QGIS or corrupts data: | No | Copied to github as #: | 29010 |
Description
Having a set of points with a 'code' attribute consisting of a 2-char code+5digits I wanted to select all dutch points (starting with NL).
Reading substr(string,start[,length]) I tried:
eg: code = "NL1234"
substring("code", 0, 2) => 'NL'
then I thought to start on position 1 so did
substring("code", 1, 2) => 'NL'
I actually started questioning because I started of with
substring("code", 0, 3) => 'NL' thinking it was from-till
then tried
substring("code", 0, 2) and also came to 'NL'
So there are three ways to find the first 2 chars :-)
Isn't this confusing? Should an expression not either be zero-based or one-based?
We could argue that expressions are not for devs/coders so we deliberately choose for 1-based. But I think then we should put that in the Help text (so peeps like me should not try zero) and using zero then should work.
But personally I would just prefer zero-based...Mmm, testing LibreOffice here, I see that using a zero as start actually returns an error)...
So this is actually a feature?
Associated revisions
fix substr help (fixes #21192)
fix substr help (fixes #21192)
(cherry picked from commit 2bed7ab0018696dba76d2c8053d22633a271486e)
History
#1
Updated by Jürgen Fischer about 5 years ago
Not reproducable here:
substr('NL1234',0,2) => 'NL'
substr('NL1234',1,2) => 'NL'
substr('NL1234',0,3) => 'NL1'
#2
Updated by Jürgen Fischer about 5 years ago
- Status changed from Open to Feedback
#3
Updated by Richard Duivenvoorde about 5 years ago
Mmm, strange. So the last example I cannot reproduce here either.
But my point that it is not clear if it is zero based or one based is still valid:
substr('NL1234',0,2) => 'NL'
substr('NL1234',1,2) => 'NL'
But we could consider this a 'feature'...
Problem would have been bigger if the second parameter would have been the end-stop instead of the length....
Close?
#4
Updated by Jürgen Fischer about 5 years ago
Richard Duivenvoorde wrote:
Mmm, strange. So the last example I cannot reproduce here either.
But my point that it is not clear if it is zero based or one based is still valid:
substr('NL1234',0,2) => 'NL'
substr('NL1234',1,2) => 'NL'But we could consider this a 'feature'...
Problem would have been bigger if the second parameter would have been the end-stop instead of the length....Close?
The help doesn't say where the indexes start - but the examples show the current behavior (ie. index start at 1 - like in postgresql for instance; postgresql however produces an empty string on substring('NL1234',0,1) and 'N' on substring('NL1234',0,2)). Except
substr('HELLO WORLD',6) → 'WORLD'
#5
Updated by Jürgen Fischer about 5 years ago
- % Done changed from 0 to 100
- Status changed from Feedback to Closed
Applied in changeset qgis|2bed7ab0018696dba76d2c8053d22633a271486e.