I'm back, after posting just a couple of hours ago about "old" dates and
getting great help.

Now I know more about what my friend is trying to do, and I'm intrigued by
the problem and am wondering how a real clever expert, like in this group,
would handle it.

She's archiving historical letters and using an Excel file. So she's got
rows with the topic, who wrote it, who received it, the date, etc. Let's
ignore the problem of "old" dates now and assume they're all after 1900.

Here's the tricky part: she's not sure about some dates.

This is classical, of course, in archival. For example:

1933? means "we think it was written in 1933 but we're not sure"

193? means "we know it was in the 1930s, but we're not sure which
year"

10/09/1934? means "we think it was that date but we're not sure"

Now, of course, you can imagine what happens when you try to put those as
dates into Excel.

Yet, she would still like to be able to sort by dates, etc. so somehow it
would be nice, and important, to preserve the "date-ness" of those date
entries.

So what to do?

It seems to me that there's no way around having more than one item per
date: The date itself, whereby you use some system to "guess" the date where
there is a question mark, and somehow some other item that records the
uncertainty in some way -- I don't know, like another column where it says
"sure/unsure" (although that isn't general enough).

Anybody have any insights? Perhaps one of you has dealt with something
similar?

Thanks,

John