My Windows 11 Regional format was previously set to Swedish. With that setup, I worked on many Qlik Cloud apps without any issues.
About a month ago, however, I noticed something strange when creating new measures: the thousand separator in number formatting started to appear incorrectly. Instead of the standard Swedish thousand separator, the character ‘�’ is used. This issue has persisted since then.
When formatting a measure, this strange separator is visible in the formatting dialog, and the same separator also appears in tables viewed by app users.
I later changed my Windows Regional format to English (UK). For new apps created after this change, the thousand separator is now correctly shown as a comma (,). However, in older apps, the original problem remains:
The incorrect ‘�’ separator is still used
The same applies to duplicated of apps with the issue.
Other Qlik developers and users experience the same behavior when they visit the infected apps, regardless of their own Regional format settings
In other words, within affected apps (including duplicates), the only available thousand separator option is ‘�’. I think the pattern suggest that it is the apps themselves that are corrupted somehow.
It may also be relevant that all our app load scripts begin with several variable settings (e.g. regional and numeric format variables).
I think the display of ‘�’ is a replacement and therefore a side effect of not being capable to resolve the set char within the used char-set and/or the font and/or the collation. This may happens on any place in the ETL as well as within the UI.
I suggest to check which char is really applied as thousand-separator. I assume it should be chr(32) but there exists various other white-spaces which may not included/supported from the char-set/font and/or from the later html/css stuff behind the rendering.
Such check might be done per copy & paste of variable-statement in an editor like Notepad++ or Excel with appropriate string-functions. Useful might be also to open the app and/or the underlying qvd-files with Notepad++ to investigate the xml meta-data of a related field.
That's not a direct solution but it should provide some insights especially by excluding certain possibilities. Beside this I would try a replacement of the separator with '+' or '#' or similar to see what happens. If it's displayed like expected? If yes you could revers it to chr(32). Maybe a back and forth ...
Thank you for your input! It led us to discover that the number formatting came from an external sharepoint imported .qvs, in which the formatting was corrupted.
.jpg){kind=link}
.jpg){kind=link}
.jpg){kind=link}