Hi all

 

I seems that there are two ways of validating an xml document with XML Schema.

 

1. Create a XMLParser object and pass it an XMLSchema object as an argument. If you use this parser object with etree.parse() or etree.fromstring(), it validates the document against the schema while parsing.

 

2. Parse the document first, and then validate it using XMLSchema.validate() or XMLSchema.assertValid().

 

I have always used the first method – I was unaware of the second one. However, I have just found a particular error that is detected by the second method, but not by the first one.

 

The error is the use of duplicate “id” attributes, if it is specified in the schema as being of type “xs:ID”. I have a document with duplicate ids. Validating with the first method passes. Validating the same document, with the same schema, using the second method, gives the following -

 

lxml.etree.DocumentInvalid: Element '{http://www.omg.org/spec/BPMN/20100524/MODEL}dataOutput', attribute 'id': 'user_row_id' is not a valid value of the atomic type 'xs:ID'., line 38

 

Is this a bug? Should I always use the second method?

 

I am using lxml version 4.1.1.

 

Thanks

 

Frank Millman

 

P.S. This is a copy of a message I posted on 26/03/2017. I got no reply to that, so I thought I would try again. I was using version 3.6.4 then, but the behaviour is the same with 4.1.1.