一个简单的解决方案是按以下方式定义您的组：

Group 1: 1-3 digits and a hyphen (-), all optional
Group 2: 2-3 digits, mandatory

以这种方式编写正则表达式将确保如果您有一个单独的2或3位数字，它将始终被识别为组2的一部分，因为组1需要识别连字符。
示例：

(\d{1,3}[-])?(\d{2,3})

通过regex101.com验证。另外，请记住适当地锚正则表达式。
如果你想让它跟踪字符串的开始和结束，你可以锚你的表达式，^表示开始，$表示结束。
示例：

^(\d{1,3}[-])?(\d{2,3})$

如果我对问题的理解正确的话，你可以尝试匹配下面的正则表达式。

^((?:(?=\d+\-)\d{1,3}|(?!\d+\-)\d{0,3}?))\-?((?:(?<=\-)\d{2,3}|(?<!\-)\d{2,3}))$

Demo
这个正则表达式可以用 * 自由间距模式 * 编写，使其具有自文档性。

(?x)            # invoke free-spacing mode
^               # match the beginning of the string
(               # begin capture group 1
  (?:           # begin a non-capture group
    (?=\d+\-)   # pos lookahead asserts later in the string is a hyphen
    \d{1,3}     # match 1 to 3 digits, as many as possible
    |           # or
    (?!\d+\-)   # neg lookahead asserts no hyphen later in the string 
    \d{0,3}?    # match 0 to 3 digits, as few as possible
  )             # end of non-capture group
)               # end of capture group 1
\-?             # optionally match a hyphen
(               # begin capture group 2
  (?:           # begin a non-capture group
    (?<=\-)     # pos lookbehind asserts prev char is a hyphen
    \d{2,3}     # match 2 to 3 digits, as many as possible
    |           # or
    (?<!\-)     # neg lookbehind asserts prev char not a hyphen
    \d{2,3}     # match 2 to 3 digits, as many as possible
  )             # end of non-capture group
)               # end of capture group 2
$               # match end of string