Regex for PascalCased words (aka camelCased with leading uppercase letter)

([A-Z][a-z0-9]+)+

Assuming English. Use appropriate character classes if you want it internationalizable. This will match words such as "This". If you want to only match words with at least two capitals, just use

([A-Z][a-z0-9]+){2,}

UPDATE: As I mentioned in a comment, a better version is:

[A-Z]([A-Z0-9]*[a-z][a-z0-9]*[A-Z]|[a-z0-9]*[A-Z][A-Z0-9]*[a-z])[A-Za-z0-9]*

It matches strings that start with an uppercase letter, contain only letters and numbers, and contain at least one lowercase letter and at least one other uppercase letter.

Adam Crume's regex is close, but won't match for example IFoo or HTTPConnection. Not sure about the others, but give this one a try:

\b[A-Z][a-z]*([A-Z][a-z]*)*\b

The same caveats as for Adam's answer regarding digits, I18N, underscores etc.

You can test it out here.

Lower camel case

this regex includes number and implements strict lower camel case as defined by the Google Java Style Guide regex validation.

[a-z]+((\d)|([A-Z0-9][a-z0-9]+))*([A-Z])?

1. The first character is lower case.
2. The following elements are either a single number or a upper case character followed by lower cases characters.
3. The last character can be an upper case one.

Here is a snippet illustrating this regex. The following elements are valid.

xmlHttpRequest
newCustomerId
innerStopwatch
supportsIpv6OnIos
youTubeImporter
youtubeImporter
affine3D

Upper camel case

Same principle as the one used for lower camel case with always a starting upper case character.

([A-Z][a-z0-9]+)((\d)|([A-Z0-9][a-z0-9]+))*([A-Z])?

Here is a snippet illustrating this regex. The following elements are valid.

XmlHttpRequest
NewCustomerId
InnerStopwatch
SupportsIpv6OnIos
YouTubeImporter
YoutubeImporter
Affine3D

The regexp that solved my problem (properly naming directories that will be recognized by FitNesse DbFit web service) is:

(^[A-Z][a-z0-9]+[A-Z]$)|(^[A-Z][a-z0-9]+([A-Z][a-z0-9]+)+$)|(^[A-Z][a-z0-9]+([A-Z][a-z0-9]+)+[A-Z]$) 

I reverse engineered these particular CamelCase rules, they are:

1. First character uppercase alpha
2. Next 1-n characters lowercase alphanumeric
3. Next character (n+1) uppercase alpha
4. Next 0 or more characters lowercase alphanumeric
No consecutive uppercase; no special characters.
Pattern may be repeated, e.g. NoChildLeftBehindSuite9102

The expression passed my testing as follows:

Camel01C is CamelCase syntax
Camel01c01 is not CamelCase syntax
Camel01C01 is CamelCase syntax
Camel01CC01 is not CamelCase syntax
Camel0a1c1 is not CamelCase syntax
Camel0a1C1 is CamelCase syntax
Camel0ac1b1C1 is CamelCase syntax
CamelC is CamelCase syntax
CamelC1 is CamelCase syntax
CamelCA is not CamelCase syntax
CamelCa1 is CamelCase syntax
CamelCa_1 is not CamelCase syntax
IbsReleaseTestVerificationRegressionSuite is CamelCase syntax
IbsReleaseTestVerificationRegressioNSuite is not CamelCase syntax
IbsReleaseTestVerificationRegressioN is CamelCase syntax