org-sort
by date will work with timestamps.
For example:
| number | date |
|--------+------------------|
| 781 | <2018-12-07 Fri> |
| 129 | <2018-12-03 Mon> |
| 321 | <2018-11-23 Fri> |
| 123 | <2018-12-04 Tue> |
With the caret in the date column type: C-c^ followed by t for sorting by time.
Wil result in:
| number | date |
|--------+------------------|
| 781 | <2018-12-07 Fri> |
| 123 | <2018-12-04 Tue> |
| 129 | <2018-12-03 Mon> |
| 321 | <2018-11-23 Fri> |
EDIT
In your case you will need to reformat those strings using regular expressions.
You can write a function such as:
(defun format-date-column (d)
(concat "<" (replace-regexp-in-string "[.]" "-" d) ">"))
and then apply it in the table formula field.
In the following example I made a third column and the function parses between columns:
#+tblfm: $3 = '(format-date-column $2)
Turning this table:
| Number | Date | timestamps |
|---------------+------------+------------|
| 1535749200000 | 31.08.2018 | |
| 1535835600000 | 01.09.2018 | |
| 1590094800000 | 21.05.2020 | |
| 1536019200000 | 04.09.2018 | |
| 1536883200000 | 14.09.2018 | |
To:
| Number | Date | datestamps |
|---------------+------------+--------------|
| 1535749200000 | 31.08.2018 | <31-08-2018> |
| 1535835600000 | 01.09.2018 | <01-09-2018> |
| 1590094800000 | 21.05.2020 | <21-05-2020> |
| 1536019200000 | 04.09.2018 | <04-09-2018> |
| 1536883200000 | 14.09.2018 | <14-09-2018> |
Which can then be sorted as explained above.