Encoding character issue??? Help!

[Mike Wilson]

I am having issue sending messages intermittently.  Apparently the vendor is sending over what appears to be an apostrophe but is not.  The Testing tool finds the character and replaces it but it wont work in the LIVE data stream.  The character in question is ”

Topic

[David Barr]

If the vendor won’t fix it you could try fixing it in TCL:

Code:

set msg [msgget $mh]
regsub -all “u2019” $msg “‘” msg
msgset $mh $msg

[Mike Wilson]

Whats the “u2019” ?

Is that the code for that specific character?

[David Barr]

Yes, I looked it up here:

http://unicodelookup.com/#0x2019/1

[Mike Wilson]

unfortunately that didn’t work either.  The sections in bold are where the issue is.

Here is what the actual text from the EMR looks like:

Comparison is made to the patient

[David Barr]

From the debug window it looks like your bad character has a hex value of hex 92, so you could use the same regsub but replace the u2019 with x92.

[Mike Wilson]

David Barr wrote:

From the debug window it looks like your bad character has a hex value of hex 92, so you could use the same regsub but replace the u2019 with x92.

David, I have tried that as well.  It simply wont recognize that character with my TCLproc.  Im stumped.

[David Barr]

What do you have the encoding set to on the inbound thread? It looks like your sender is using “windows-1252”.

[Mike Wilson]

the Vendor is Meditech C/S 5.66.  I have the encoding set to bypass right now.  I have been changing it to see if I can get any change with no luck.

[David Barr]

I’d try setting the inbound encoding to windows-1252, then using the regsub with u2019. Setting the encoding to 1252 should allow Cloverleaf to properly convert the characters to Unicode, then you can use the Unicode character code to replace with the plain ASCII equivalent.

[Mike Wilson]

David Barr wrote:

I’d try setting the inbound encoding to windows-1252, then using the regsub with u2019. Setting the encoding to 1252 should allow Cloverleaf to properly convert the characters to Unicode, then you can use the Unicode character code to replace with the plain ASCII equivalent.

That did the trick!  What a pain.  Thank you David!

[Author]

Replies