March 05 1982 was not a monday, but a friday.
But the line below will convert it correctly anyway:
date -d 'Mon Mar 05 18:34:38 1982' +%m%d%Y
It should be possible with sed
, but it would be a tedious task. Not recommended for sure (in this case, handling date conversion).
If the date fields are always at the start of the lines, you don't even need to search for them: just run "date" on the first 24 chars. If that's invalid, "date" will return non-zero exit code.
If the dates are in certain field number(s) of the csv file, I'd use awk. Have a look at the strftime() function of awk for conversion (this will be similar to the "date" example above).
If the dates can be anywhere in the lines of the CSV file, use grep to find the date strings, and convert them with "date". (see above)