Originally posted by: SystemAdmin

First, if there is any documentation or training that anyone knows of, please let me know, cause I have a million questions.

A couple of them, for anyone who has mapped different character sets, are;

I have Chinese traditional, UTF8 on the input, mapping to an X12 215, N1 loop (address).

Thinking I had to, I change the X12 type tree to be the same character set.

The country code in English is TW for Taiwan, 2 characters, 6 bytes long.

When I map it to the X12 215 country code, defined as 2-3 bytes in the type tree, even though I said the output is UTF8, I only get one character, 3 bytes.

Am I doing something wrong, or is this an extent on the functionality?

I know newer version of Oracle have different functions; Size (for byte size) and Charsize( for the number of characters). I guess I was hoping WTX would have this also.
Thank you.
#DataExchange
#IBM-Websphere-Transformation-Extender
#IBMSterlingTransformationExtender

Originally posted by: SystemAdmin

Re-evaluating my post, I believe it’s not applicable. I'm still trying to wrap my head around this whole concept.

The X12 country code field is 2 min and 3 max bytes in length, and I’m trying to put 6 in there, even though the visible, human readable representation is only two characters. So getting the one character would be correct.

I did do a bit of eval that I thought I would post. If anyone has worked with multi byte data and has ANY input on ANY aspect of it, I would apprecaite any information.
Using the LEFT function in WTX, where you specify the field and the number of bytes/characters from the left you want, I was wondering whether it would take the X number of characters or bytes from the left.

So going to and from a WTX defined Chinese Trad/UTF-8 fields, when I did LEFT(country code, 1), I got the first one character, equating to 3 bytes. When I did LEFT(country code, 2), I get the two characters, equating to 6 bytes.
I then decided to try a bunch of combinations, using the 6 byte/2 character country code, results in Hex:

<hr />
western to western, COUNTRY code
in
E5 8F B0 E7 81 A3
out
E5 8F B0 E7 81 A3<hr />
western/native to Chinese:
in
E5 8F B0 E7 81 A3
out
C3 A5 8F C2 B0 C3 A7 C2 81 C2 A3<hr />
Chinese to western

NO OUTPUT

<hr />
Using the left function, changing the defintion of the fields:

Chinese to Chinese:

LEFT(country code, 1) = E5 8F B0

western to Chinese

LEFT(country code, 1) = C3 A5

Chinese to Western

LEFT(country code, 1) = NO OUTPUT.
All in all, I think these are better results then I expected. When attempting to move a multi byte Chinese character with no representation in the the western code set, to a western defined field, I got no output, even when I coded "Move the 1 left most character from the Chinese string to the western defined field". It would seem the map knew the code set.
#DataExchange
#IBMSterlingTransformationExtender
#IBM-Websphere-Transformation-Extender

Originally posted by: SystemAdmin

In WTX 8.2, there is a new feature, where you can define size in characters or bytes.
#DataExchange
#IBMSterlingTransformationExtender
#IBM-Websphere-Transformation-Extender

Originally posted by: jvanboga

one option is to convert the data before using. I had an htm/xml file that was created using an India character set. I mapped it to an ASCII htm/xml set then processed data from there. Just needed to define the correct character sets.
#IBM-Websphere-Transformation-Extender
#IBMSterlingTransformationExtender
#DataExchange

Originally posted by: SystemAdmin

Thanks for the input.

Skd, of course, we’re moving to 8.1 currently. Can I ask; where you define the size in bytes or characters, is that where in the earlier versions you normally define the min and max size? Is it kind of the same, only there’s an option for byte or characters, kind of like under Item Subclass, the option for character or binary?
Jvanboga, It seems when I save the document to ASCII (128), it turns the characters into question marks.

After doing even more research, it looks like EDIFACT 4 is Unicode compliant, but it doesn’t look like the type trees we have are. In example, the country code is defined as 1-3 bytes, and will not fit a 6 bytes, 2 Chinese character country code.
#DataExchange
#IBM-Websphere-Transformation-Extender
#IBMSterlingTransformationExtender