Re: How to declare an element of type xs:unsignedInt, restricted to one to five digit characters?

[Sloane, Brandon]

The xs:unsignedInt is still providing xsd validation (so someone cannot unparse 
a value of "a"). If you wanted to take full advantage of this, you need a more 
specific xsd type in a maxValue set. Normally Daffodil does not pay attention 
to xsd restrictions during parse/unparse (although it can optionally do xsd 
validation before/after a parse/unparse). If you want to take avantage of xsd 
restrictions during parse/unparse, you can use the dfdl:checkConstraints() 
function inside of an assert annotation.


Note that setting a max value of 99999 and adding an assert is not quite the 
same thing using lengthPattern. Using maxValue+assert would fail to parse the 
input "123456". Using dfdl:lengthPattern would parse "12345" and leave "6" in 
the dataStream for the next element to parse. Depending on your format, one of 
these behaviours might be preferable.


Often, what works best is to parse all of "123456", then, after parsing is 
complete, fail the resulting infoset using standard XSD validation. This would 
be done with a standard xsd restriction (and no dfdl assertions) and 
lengthKind=delimited and parsing with validation on (or using an external 
validator). This allows for parsing syntactically valid data, which may make 
debugging semantically invalid data easier. This is not an option succesfully 
parsing requires backtracking when the integer is too big.

________________________________
From: Costello, Roger L. <coste...@mitre.org>
Sent: Wednesday, June 19, 2019 12:51:12 PM
To: users@daffodil.apache.org
Subject: Re: How to declare an element of type xs:unsignedInt, restricted to 
one to five digit characters?


The disadvantage that I can see with that solution is the xs:unsignedInt is 
doing essentially nothing – it may as well be xs:string. The pattern is doing 
all the constraining. Do you agree?



Thanks Brandon.



/Roger



From: Sloane, Brandon <bslo...@tresys.com>
Sent: Wednesday, June 19, 2019 12:13 PM
To: users@daffodil.apache.org
Subject: [EXT] Re: How to declare an element of type xs:unsignedInt, restricted 
to one to five digit characters?



Assuming your input is plaintext, you could use the dfdl:lengthPattern 
attribute:



            <xs:element name="value"
                 type="xs:unsignedInt"
                 dfdl:lengthKind="pattern"
                 dfdl:lengthPattern="[0-9]{1,5}"
                 />



You could also use lengthKind="delimited" and add an assertion (and/or xsd 
restriction) on the value.

________________________________

From: Costello, Roger L. <coste...@mitre.org<mailto:coste...@mitre.org>>
Sent: Wednesday, June 19, 2019 9:30:11 AM
To: users@daffodil.apache.org<mailto:users@daffodil.apache.org>
Subject: How to declare an element of type xs:unsignedInt, restricted to one to 
five digit characters?



Hello DFDL community,

The below DFDL schema says that the input must contain a string that represents 
an unsignedInt and the string must not contain more than five digit characters.

Eek!

That's not correct. The below DFDL schema says that the input must contain a 
string that represents an unsignedInt and the string must contain *exactly* 
five digit characters.

Is there a way to write the DFDL schema to specify that the input must contain 
a string that represents an unsignedInt and the string must contain *one to 
five* digit characters? That is, how to specify that the input could be any of 
these strings:

"0", "1", ..., "99999"

<xs:element name="input">
    <xs:complexType>
        <xs:sequence>
            <xs:element name="value"
                 type="xs:unsignedInt"
                 dfdl:lengthKind="explicit"
                 dfdl:length="5"
                 dfdl:lengthUnits="characters" />
        </xs:sequence>
    </xs:complexType>
</xs:element>