UltraEdit is a text editor. It cannot process binary data stored in a *.xls or *.xlsx file. It can process data stored in a comma-separated values file with using a comma , or a semicolon ; or a vertical bar (pipe) | or a horizontal tab character as separator and having usually the file extensions .csv or .tsv. Such files are not Excel files. Excel is just one of many applications which can read in files with comma-separated values (not really 100% correct regarding to multi-line values).

I do not understand what you really want. Please post an example of the file on being opened in UltraEdit and what you want as result of running one or more Perl regular expression replaces. Three to five lines should be enough to explain how the data should be processed by the regular expression replace(s).

Please use the last but one item (second from right side) on the toolbar above the edit field in your browser to format the selected input and the selected output example lines in your next post as code block as I have done on your initial post.

1. Make copy of your data
2. Find what: 3. Replace with: 4. Mark Regular expressions: and choose Perl
5. Mark Replace all is from top of file (if needed)
6. Click Replace all

The Perl regular expression find/replace as posted by Ovg works fine using a backreference.

Another possible Perl regular expression replace is using as search string (?=\d\d\>) and a single dot . as replace string. The search expression uses a positive look-ahead to check on each character in the file if there are next two digits at end of a word without selecting them. If this condition is true, the replace is done by inserting the decimal point.

Run first a Perl regular expression with search string \<(\d\d)\> and replace string 0.\1 and run next a second Perl regular expression replace with \d\K(\d\d)\> as search string and .\1 as replace string.

The result for the tab separated input example data is:
The search string  \<(\d\d)\> means:

\< ... find a string at beginning of a word. Space, horizontal tab, line-feed, hyphen, dot are not word characters. So a word is for the Perl regular expression just a series of digits for the input data.

(...) ... mark the string found by the expression inside the first pair of round brackets for being back-referenced in search or replace string with \1 as done in the replace string to keep the two digits.

\d\d ... find two digits.

\> ... the two digits must be at end of a word.

In other words this search expression matches only numbers with exactly two digits. The minus sign (in real the hyphen character) is not a word character.
he replace string inserts the digit 0 and a dot left to the two found digits which are kept due to back-referencing them.

The second search expression \d\K(\d\d)\> means:

\d ... find a digit.

\K ... keep back the matched digit which means unselect the already matched digit and start selecting found string from this position in character stream.

(...) ... mark the string found by the expression inside the first pair of round brackets for being back-referenced in search or replace string with \1 as done in the replace string to keep the two digits.

\d\d ... find two digits.

\> ... the two digits must be at end of a word for a positive match.

In other words the search expression searches for three digits at end of a word, ignores the first digit for the replace and inserts a dot left the two found digits which are kept due to back-referencing them.

There can be used the Perl regular expression: \<(?:\d+,)+\d+\.\d+\>

\< ... find a string at beginning of a word.

(?:...) ... defines a non-marking group.

\d+, ... find a digit one or more times followed by a comma.

+ ... the non-marking group must be applied one or more times.

\d+ ... find a digit one or more times

\.\d+ ... find next a dot which must be escaped with a backslash to be interpreted as literal character and one or more digits.

\> ... the last digit must be at end of a word.

The expression matches also a value like 1,234,567.89.