Can you stop StrConv(string, vbProperCase) from removing accents?

[haggis999]

One of my Access forms has a textbox that is coded to automatically convert all text input to proper case, using StrConv(string, vbProperCase). I have just noticed that some accented characters in the original text are replaced by non-accented versions of those characters, while other accented characters survive the process. For example, a c with a circumflex accent is stripped of its accent, while a c with a cedilla is unaffected (both these accents are used in French, Portuguese and Turkish, in addition to other languages).



If I disable the StrConv code then everything works fine. The problem exists in both Access 2007 and Access 2016.

Is this behaviour a bug or by design? More importantly, is there a way to persuade StrConv to stop messing with the accents?

[ranman256]

I doubt it. StrConv is a built in function.

[isladogs]

I think all you can do is a work round like this:
a) go through & identify all characters that get modified using StrConv
b) before running StrConv - replace with a unlikely symbol like ¬ or a string like CFLX
c) run StrConv
d) reverse replace all instances of CFLX or ¬ or whatever

It will work ... but be a bit tedious to setup

[haggis999]

I think all you can do is a work round like this:
a) go through & identify all characters that get modified using StrConv
b) before running StrConv - replace with a unlikely symbol like ¬ or a string like CFLX
c) run StrConv
d) reverse replace all instances of CFLX or ¬ or whatever

It will work ... but be a bit tedious to setup

Thanks for that suggestion. It would do the job, but this problem is only a minor issue in the app where I discovered this odd behaviour, so I might not bother to fix it. My question was simply driven by curiosity. I was struggling to think of any logical reason why StrConv handled some accented characters and not others.

I have since looked a little closer at StrConv's optional LocaleID parameter. This lets you use something other than the default LocaleID on your system, but I suspect that changing the LocaleID (and thus the code page) would simply change which accented characters didn't work properly.

[orange]

You may be able to work with codepage --this uses transfertext