Regex Replace cannot lock on to correct character offset in replacement

Can someone help me figure out the following:

if true
  myStr = 'margins1.1.5 Configuration management1.1.6 Cool engineering1.1.7 Modeling';
  expression = '[^\d^ ]\d(\.\d){1,}';
  replace_0 = ' [$0]';
  replace_1 = ' [$1]';

    newStr_0 = regexprep(myStr,expression,replace_0)
    newStr_1 = regexprep(myStr,expression,replace_1)

    desiredStr = 'margins [1.1.5] Configuration management [1.1.6] Cool engineering [1.1.7] Modeling'
  end

I am getting the following and want to get desiredStr

    newStr_0 =

    'margin [s1.1.5] Configuration managemen [t1.1.6] Cool engineerin [g1.1.7] Modeling'

newStr_1 =

    'margin [.1.5] Configuration managemen [.1.6] Cool engineerin [.1.7] Modeling'

desiredStr =

    'margins [1.1.5] Configuration management [1.1.6] Cool engineering [1.1.7] Modeling'

Answer 1

[Guillaume]

I don't understand why you have two consecutive replacements.

This will give you the correct result:

regexprep(myStr, '(\d\.)+\d', ' [$0]')

Your regex doesn't make much sense, [^\d^ ] means match any character that do not match \ or d or ^ or a space. Note that escapes don't work in [] and the ^ has only special meaning as the first character inside []. edit: brain went mushy. See below for correct interpretation.

edit: Note that the above will only match numbers from 1 to 9, so not 1.1.10. if the latter is desired then use \d+ instead of \d.

[Walter]

That won't work for the case: 'dated September 25, 2000.3.2 Exhibit 2,'

it will give you: 'dated September 25, 200 [0.3.2] Exhibit 2,'

hence the extra trickery.

[Guillaume]

As per my edit, if you want it to match multiple digit numbers replace \d by \d+.

If you want to avoid adding an extra space when one already exists:

regexprep(myStr, ' ?((\d+\.)+\d+)', ' [$1]')

[Stephen23]

"Note that escapes don't work in [] "

According to what? This escaped character works:

>> regexp('abcd2345','[^\d]+','match')
ans =
  'abcd'

The regular expression help also includes this example: " \S Any non-white-space character; equivalent to [^\f\n\r\t\v]". It is not clear how these show that escaped characters do not work inside [].

[Walter Roberson]

Yup and you can need to escape ] characters inside a []. There is a special convention about ] immediately after [ or immediately after [^ but those conventions are not enough to remove the need to escape ] characters inside []. What is true is that quantifiers and dot and () lose special meaning inside []

[Guillaume]

Yes, not sure what I was on about. Of course, escapes work in []. Still, the second ^ in the [] is a literal ^ not a do not match next character indicator. So the [^\d^ ] means do not match any of digit, ^ or space.