|
|
 next
|
  Subject: EDI (X12 or EDIFACT) to XML questions re: Encoding & Schema
Author: Minollo I.
Date: 16 Dec 2005 05:48 PM
|
>1. In the samples, (demo.java) when converting a CSV to XML, the XML Output has
>a prolog that declares the encoding
>
><?xml version="1.0" encoding="utf-8"?>
>
>I have changed the sample to do EDI X12 to XML by
>
>FileInputStream fis = new
>FileInputStream( "./831.x12");
>StylusFile in_1 = sff.createStylusFile("adapter:EDI:newline=crlf:val=yes:decode=yes:field=yes:len=no:seg=yes:tbl=yes:typ=yes:opt=no:eol=yes", fis);
>
>The output XML for EDI doesn't contain prolog? Why?
I would think the underlying reason is that the EDI converter always generates output in UTF-8 format (converting from the proper EDI message format). XML documents with no PI are assumed to be UTF-8 encoded.
But I tend to agree with you; it would be a cleaner approach to output the PI even if we know the output is always UTF-8.
>2. In addition, is there a way to associate a schema with the Output XML ?
You mean automatically creating a schema for the XML format of the EDI message (like the File > Document Wizard... > X12/EDIFACT to XML Schema wizards do), and automatically adding a reference to it in the generated XML?
No, we don't support that capability; you can clearly do that from inside Stylus Studio in a couple of steps, but to achieve that from the API we would have to add a couple of methods to let you create and assign the schema automatically.
I'm filing change requests about both topics, as they both sound like good points. #1 is a nuisance, and can be addressed very quickly; #2 will require a bit more of thinking.
Thanks,
Minollo
|
|
next
|
  Subject: EDI (X12 or EDIFACT) to XML questions re: Encoding & Schema
Author: toadie d.
Date: 10 Jan 2006 10:29 AM
|
Actually, come to think of it. If you implement #1 without #2 , that's probably even worse.
Currently, via the API layer, the caller controls the creation of output stream. We can pre-insert the prolog and necessary doctype.
prior to handing th stream to the adapter.
If you implement #1 without #2, that makes it even more difficult to
implement the XSD/DocType work around.
Thanks much.
Toadie
>>1. In the samples,
>(demo.java) when converting a
>CSV to XML, the XML Output has
>>a prolog that declares the
>encoding
>>
>><?xml version="1.0"
>encoding="utf-8"?>
>>
>>I have changed the sample
>to do EDI X12 to XML by
>>
>>FileInputStream fis = new
>>FileInputStream(
>"./831.x12");
>>StylusFile in_1 =
>sff.createStylusFile("adapter:
>EDI:newline=crlf:val=yes:decod
>e=yes:field=yes:len=no:seg=yes
>:tbl=yes:typ=yes:opt=no:eol=ye
>s", fis);
>>
>>The output XML for EDI
>doesn't contain prolog? Why?
>
>I would think the underlying
>reason is that the EDI
>converter always generates
>output in UTF-8 format
>(converting from the proper
>EDI message format). XML
>documents with no PI are
>assumed to be UTF-8 encoded.
>
>But I tend to agree with you;
>it would be a cleaner approach
>to output the PI even if we
>know the output is always
>UTF-8.
>
>>2. In addition, is there a
>way to associate a schema with
>the Output XML ?
>
>You mean automatically
>creating a schema for the XML
>format of the EDI message
>(like the File > Document
>Wizard... > X12/EDIFACT to
>XML Schema wizards do), and
>automatically adding a
>reference to it in the
>generated XML?
>No, we don't support that
>capability; you can clearly do
>that from inside Stylus Studio
>in a couple of steps, but to
>achieve that from the API we
>would have to add a couple of
>methods to let you create and
>assign the schema
>automatically.
>
>I'm filing change requests
>about both topics, as they
>both sound like good points.
>#1 is a nuisance, and can be
>addressed very quickly; #2
>will require a bit more of
>thinking.
>
>Thanks,
>Minollo
|
|
|
|
|
top