Question

I am working on creating an XSD for a web service that will take in an ID number as an element in the XML. These ID numbers consist of 10 consecutive digits ([0-9]{10}), but I was trying to create a regular expression that could exclude certain elements from this range.

For example, here is the restriction I have currently in my XSD:

    <xsd:restriction base="xsd:string">
        <xsd:pattern value="[0-9]{10}" />
    </xsd:restriction>  

I need the restriction to allow a string of [0-9]{10} that doesn’t fit the following IDs:

    All 0's:         [0]{10}
    Starting with 6: [6][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]
    Starting with 000: [0][0][0][0-9][0-9][0-9][0-9][0-9][0-9][0-9]
    Starting with 999: [9][9][9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]
    Ends with 2 0's: [0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0][0]
    4 0's in Middle: [0-9][0-9][0-9][0][0][0][0][0-9][0-9][0-9]

Is this possible to do from within the XSD or regular expression?

Thanks.

Answer 1

I’d rephrase your restrictions a bit:

• The first digit must not be a 6.
• At least one of the last two digits must not be a zero.
• At least one of the middle four digits must not be a zero.

The first restriction, an ID only consisting of zeroes, is actually included in the two last restrictions.

The first restriction can be expressed by a set of allowed characters that does not include 6, i.e. [0-57-9].

For the other restrictions, a straightforward solution is to start at the beginning of a section that must not consist only of zeroes and assume a non-zero digit; if that assumption is true, the remaining digits may include zeroes; otherwise the first digit in that section must be a zero and for the remaining characters, this rule can be repeated recursively until only one character is left: ([1-9][0-9]{3}|0(... repeat for three digits, then two digits, ...))

Therefore, a suitable RegEx would be:

[0-57-9][0-9]{2}([1-9][0-9]{3}|0([1-9][0-9]{2}|0([1-9][0-9]|0[1-9])))[0-9]([1-9][0-9]|0[1-9])

Update: The additional restrictions require the following:

• At least one of the first three digits must not be a 0.
• At least one of the first three digits must not be a 9.

This can be included the same way as above, accepting either anything except 0 and 9, or either of these two numbers:

([1-57-8][0-9]{2}|0([1-9][0-9]|[0-9][1-9])|9([0-8][0-9]|[0-9][0-8]))([1-9][0-9]{3}|0([1-9][0-9]{2}|0([1-9][0-9]|0[1-9])))[0-9]([1-9][0-9]|0[1-9])

The new part is in the front of the expression:

([1-57-8][0-9]{2}|0([1-9][0-9]|[0-9][1-9])|9([0-8][0-9]|[0-9][0-8]))

So,

• either the ID starts with neither a 0 nor with a 9. In that case, there are no restrictions for the next two digits.
• or the ID starts with a 0. In that case, one of the next two digits must not be a zero, either the first one or the second one.
• or the ID starts with a 9. In that case, one of the next two digits must not be a nine, either the first one or the second one.