Policy Guide › Data Lookup › Counting Unique Domains

Counting Unique Domains

An important feature of Message Attribute (msgattr) lookup is the ability to count the number of unique domains in a list of e-mail recipients. This enables administrators to block users from sending individual e-mails to, for example, more than five companies at a time.

This feature relies on CA DataMinder successfully extracting the domain element of an SMTP e-mail address. The domain element comprises either two segments ('short domain') or three segments ('long domain') after the '@' symbol. For example:

• Short domain: spencerrimmel@unipraxis.com
• Long domain: spencerrimmel@unipraxis.co.uk

When extracting the domain from an SMTP e-mail address, CA DataMinder always assumes this is a short domain and so extracts the final address two segments unless these final two segments match any one of a list of known exceptions, in which case CA DataMinder infers that is must extract a long domain (that is, the final three segments).

Long domain example

Consider an e-mail sent to two recipients, lsteel@unipraxis.co.uk and srimmel@monitrax.co.uk.

It is clear from the recipient addresses that the e-mail is being sent to two companies, Unipraxis and Monitrax. But if CA DataMinder did not have the ability to extract long domains, it would only extract the final two address segments, 'co.uk', and so would incorrectly infer that the domain was the same for both recipients (giving a unique domain count of 1).

However, because '.*.uk' is one of the known domain exceptions, when CA DataMinder detects '.co.uk' in an e-mail address it recognizes this as part of a long domain and so extracts the final three segments. In the example above, this enables CA DataMinder to recognize that the two recipients belong to two different companies and so it correctly calculates the unique domain count to be 2.

List of known long domains

This list of known exceptions is hard-coded within CA DataMinder and includes all commonly used long domain patterns such as '.*.uk' and '.ru.com'. However, as the list of worldwide domains continues to grow, it is inevitable that new long domains will emerge that do not match this hard-coded list.

If required, you can also supplement this list in the user policy; to do this, edit the Additional Long Domain Endings setting.

Adding to the list of long domains

If you do need to supplement the default list of long domain patterns, you need to edit the Additional Long Domain Endings user policy setting; find this in the Definitions policy folder. When you edit this setting, be aware of the following:

• All entries must start with a period and contain exactly two periods, for example: .ins.kr .fin.tz
• If required, you can use a * wild card in place of an entire segment, for example: .*.pb

However, we recommend you avoid using wildcards where possible because the scope of the resulting match may be greater than anticipated, causing in short domains to be inadvertently treated as long domains. This is particularly true if the domain element of an address includes a geographical subdomain. For example, adding '.*.pb' to the exceptions list would cause CA DataMinder to treat ny.unipraxis.pb and london.unipraxis.pb as separate domains.