How to specify metadata of XML file in PROFILING

venki · Post by **venki** » Mon Feb 23, 2009 7:03 am

--------------------------------------------------------------------------------

Hi,

I have another reqirement. Now Source file is XML.
How to read XML metadata into PROFILING.

I have an idea of reading flat files, for this we have created .txt and QETXT.INI files in particular folder.
Like that how to read XML metadta.

Can some one help me....

Thanks in advance.

eostic · Post by **eostic** » Mon Feb 23, 2009 9:45 am

When you say "profiling," do you mean using "Information Analyzer?" ...if so, IA uses ODBC, so you would need a method to read XML via ODBC.......I've seen a few interesting methods here, including Federation Server, Data Direct and a few others........ however, if you have DS, I would expect you to be most successful by reading the xml structure with DS, picking out the specific bits that you want to profile, adjusting things manually as necessary (because XML comes in many flavors and is not conducive to ODBC), and putting them into a structure that you can more easily profile....

...if you are not talking about Information Analyzer, please be more specific.

Ernie

venki · Post by **venki** » Tue Mar 17, 2009 5:11 am

I am using Information Analyzer (V8.0) only.

Can you please explain how to specify XML metadata into QETXT.INI file.

I have knowledge to specify fixed width files metadata (fileds) into QETXT.INI and import same into IA.

Can any one please assist me on this?

Thanks in advance

Regards,

Venkat.

eostic · Post by **eostic** » Tue Mar 17, 2009 6:20 am

I'm not exactly sure what this .ini is, but I suspect that it is part of perhaps the DataDirect ODBC driver set? I have not personally used the DataDirect ODBC Driver for XML.....but have discussed it with others who have had difficulty using it. I'm sure it works, but it will probably require that you normalize the XML document, so pick and choose for simplicity only a single repeating node that will give you a set of rows. A complex document will probably need "many" tables to be defined.

At quick glance with a search engine, there are lots of resources on the web that talk about DataDirect's ODBC driver for XML. Let us know how it goes. In the meantime, if you decide to read it first with DataStage, or to define it with Federation Server, we'll be here.

Ernie

vmcburney · Post by **vmcburney** » Tue Mar 17, 2009 5:55 pm

Do not try to define an XML file using QETXT as it's for delimited text files. You are much better off flattening the XML structure with DataStage into a database table. Info Analyser profiling works by looking at data as tables and columns - it will not handle the heirarchy of an XML file.